Remote control method for mine trackless rubber-tyred vehicle based on 5G communication

Abstract

The invention relates to the field of mine transportation, and discloses a mine trackless rubber-tyred vehicle remote control method based on 5G communication, which comprises the following steps: constructing a dynamic beam forming model oriented to a mine complex scene, sensing environment change in real time, adjusting a communication direction, combining a slicing and resource scheduling mechanism of a 5G network, and dynamically optimizing a time slot resource allocation strategy. And establishing a coupling model of a communication link state and a vehicle control system, and realizing cooperative scheduling and adaptive adjustment of communication parameters and control instructions. By constructing a high-reliability and low-delay 5G communication link, stable remote control of the vehicle in severe mine environments such as high temperature, high humidity, multiple obstacles and the like is realized, and the problems of communication blocking, signal interruption, slow control response, poor system linkage and the like in the prior art are effectively solved; and the safety, the intelligent level and the operation efficiency of mine transportation operation can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of mine transportation, and specifically to a remote control method for trackless rubber-tyred vehicles in mines based on 5G communication. Background Art

[0002] In the mine environment, the remote control of trackless rubber-tyred vehicles has always faced severe challenges in signal transmission. Existing technologies mostly rely on traditional communication methods and usually adopt static beamforming technology. However, in a complex mine environment, such as curved roads, narrow roadways or areas with multiple obstacles, this technology cannot adjust the beam direction in a timely manner, resulting in communication interruption and signal attenuation. Under traditional methods, the stability of signals is difficult to guarantee, greatly affecting the real-time performance and reliability of vehicle control.

[0003] Existing time slot resource allocation methods usually rely on fixed priority queues and bandwidth ratio allocation strategies. In some applications with high real-time requirements, there are often situations where the priority data stream cannot obtain resources in a timely manner, resulting in delays in control data. This fixed bandwidth allocation mode not only fails to effectively cope with network congestion but also leads to slow response of control instructions in emergency situations, affecting the safety of mine operations.

[0004] In addition, most traditional communication and control systems work independently, lacking dynamic feedback and coupling of communication status and vehicle status. This makes the optimization of communication links unable to follow the dynamic changes of vehicles in real time, resulting in a mismatch between the communication network and vehicle behavior, and even the phenomenon of "information loss". The designs of existing technologies do not consider the impact of environmental changes on communication parameters, so the system has poor adaptability and insufficient flexibility in emergency situations.

[0005] Therefore, the present invention proposes a remote control method for trackless rubber-tyred vehicles in mines based on 5G communication to solve the deficiencies of existing technologies. Summary of the Invention

[0006] Aiming at the deficiencies of existing technologies, the present invention provides a remote control method for trackless rubber-tyred vehicles in mines based on 5G communication, which solves the problems of unstable signal transmission, poor real-time performance and slow response of control instructions in existing methods by dynamically adjusting communication beams, time slot resource allocation and the coupling of communication and control systems.

[0007] To achieve the above objectives, the present invention is realized through the following technical solutions: A remote control method for trackless rubber-tyred vehicles in mines based on 5G communication, including the following steps: S1. The vehicle completes identity authentication through the quantum key distribution protocol and generates a quantum key, and dynamically accesses the 5G-TSN network based on the authentication result. The 5G-TSN network is configured as a distributed communication framework supporting federated learning; S2. Based on the quantum key, the vehicle collects roadway environment data in real time, encrypts it to generate an encrypted data packet, and uploads it to the edge server through the established 5G-TSN network; S3. Based on the encrypted data packet uploaded in S2, the edge server constructs a digital twin model to predict the dynamic changes of the environment. Meanwhile, a global collaborative control strategy is generated through federated learning; S4. Execute the motion instruction based on the global collaborative control strategy, and trigger the dynamic adjustment mechanism using the roadway environment data to update the vehicle steering and speed parameters; S5. According to the feedback data output by the dynamic adjustment mechanism, optimize the beamforming parameters of 5G communication and the TSN time slot allocation. Meanwhile, encrypt and store the operation logs through the blockchain to form a closed-loop audit link.

[0008] Preferably, the execution of the quantum key distribution protocol in step S1 includes: Generate an encryption key through an improved BB84 protocol. Among them, the vehicle and the security authentication center exchange polarization-encoded quantum states, negotiate to generate a key after screening matching bases, and use the decoy state detection to resist the photon number splitting attack; The security authentication center distributes the public key to the vehicle for subsequent data integrity verification.

[0009] Preferably, the dynamic access to the 5G-TSN network in step S1 includes: Encrypt the communication request based on the generated quantum key, and dynamically allocate TSN time slots according to the real-time signal quality parameters. The time slot priority is calculated through the following formula: ; where, is the vehicle spacing, is the roadway curvature, is the signal-to-noise ratio; identifies the priority score or selection probability of the candidate path , and the larger the value, the higher the priority; the weight coefficient is optimized online through a reinforcement learning algorithm, and the optimization goal is to minimize the weighted sum of the end-to-end communication delay and packet loss rate.

[0010] Preferably, the roadway environment data in step S2 includes: The three-dimensional point cloud data collected by LiDAR, the vehicle pose obtained by the UWB positioning module , and the acceleration information of the inertial navigation unit; Before data encryption, the public key is required to verify the data integrity. The verification method is to calculate the data hash value and compare it with the encrypted hash value.

[0011] Preferably, when uploading encrypted data in the 5G-TSN network in step S1: Divide the priority queue according to the data type, where the point cloud data is assigned the highest priority, the pose data is assigned the second priority, and the acceleration data is assigned the lowest priority; The bandwidth allocation ratio of the priority queue is dynamically adjusted by the following formula: ; where, are the real-time queue lengths of the point cloud and pose data respectively; are the bandwidth occupancy ratios of the point cloud, pose, and acceleration data respectively.

[0012] Preferably, the specific process of constructing the digital twin model in step S3 includes: Perform voxel grid downsampling on the encrypted point cloud data uploaded by multiple vehicles to extract the key feature point set ; Based on the feature point set Construct a three-dimensional grid map And through The network predicts the roadway deformation parameters within the future ; ; The updated global model parameters are sent to each vehicle for iterative training until the model converges.

[0013] Preferably, the federated learning process in step S3 includes: The edge server distributes the initial global model parameters to each vehicle, and each vehicle performs local training based on the local roadway environment data and encrypts and uploads the model parameters to the edge server; The edge server calculates the aggregation weight dynamically according to the Euclidean distance between the vehicle and the collapse risk area , and generates the updated global model parameters through weighted averaging: ; where, is the weight of the global model in the th iteration, which represents the global model weight merged from the weights of each local model in the current iteration; is the local model parameter of the nd vehicle in the th round; is the smoothing coefficient, is the total number of participating vehicles; The updated global model parameters are sent to each vehicle for iterative training until the model converges. ​

[0014] Preferably, the dynamic adjustment mechanism in the step S4 includes: Based on the vehicle spacing monitored in real time and the roadway curvature , calculate the steering angle adjustment amount and the acceleration adjustment amount through the following formula: ; wherein, is the preset safety distance, is the reference speed, is the weight coefficient determined by offline calibration, is the current real-time speed of the vehicle.

[0015] Preferably, the vehicle steering and speed parameters in the step S4 include the steering angle adjustment amount and the acceleration adjustment amount , and the specific steps include: Input and into the vehicle bottom controller, and update the actual steering angle and the acceleration through the following closed-loop control law: ; wherein, is the state variable in the th iteration, representing the new state updated based on the state of the previous time step at the current time step; is the state variable in the th iteration, representing the state of the device at the previous time step; is the amount of change of the state variable in the th iteration, representing the new state update amount; is the auxiliary variable associated with the state variable , representing a certain state or input information in the th iteration; is the change amount of the variable in the state update, representing the increment of to during the iteration process from ; is the control period, and satisfies .

[0016] Preferably, the steps for optimizing the beamforming parameters of 5G communication in the step S5 include: According to the vehicle position output by the dynamic adjustment mechanism and the collapse risk area Prediction boundary distance , calculate the beam direction angle adjustment amount , the formula is: ; Among them, is the current vehicle speed (unit: m / s); is the current vehicle heading angle; is the predicted distance from the vehicle position to the boundary of the collapse risk area (unit: m); Update the current beam direction angle based on the adjustment amount: ; Among them, represents the beam parameter in the th iteration; represents the beam parameter in the th iteration; represents the learning rate or adjustment factor, used to control the step size of each iteration; represents the change amount of the beam parameter in the current step, dynamically set according to the channel state information to .

[0017] The present invention provides a remote control method for a mine trackless rubber-tyred vehicle based on 5G communication, which has the following beneficial effects: 1. The present invention introduces a dynamic beamforming mechanism combined with a roadway environment prediction model to realize the autonomous direction adjustment of the communication link in complex mine scenarios. Compared with the traditional static beam control method, there is a problem of "blind area communication interruption" in tunnel bends or occlusion areas, which helps to improve the communication continuity of the vehicle in low-signal areas and solves the problem that the channel direction is difficult to match in real time.

[0018] 2. The present invention adopts a TSN priority queue dynamic bandwidth allocation method to adjust the network occupancy ratio of various sensing data according to the queue length. This mechanism is different from the conventional fixed bandwidth allocation strategy and can dynamically bias towards key data streams in scenarios with limited network resources, solving the problems of high transmission delay and control lag of pose information in data congestion in existing methods.

[0019] 3. The present invention introduces a joint adjustment mechanism of vehicle feedback control and 5G communication parameters to establish an end-to-end closed-loop dynamic control link, which is different from the traditional "control-communication" decoupled design. This method directly applies the vehicle state to the communication parameter adjustment, solving the problems of slow response and untimely adjustment of communication strategies in existing technologies, and making the system more adaptive.

[0020] 4. The present invention integrates a blockchain encryption audit mechanism to seal communication and control logs in real time. This method provides a traceable solution that does not rely on a central server in a mine operation environment, avoiding problems such as easy tampering and low query efficiency of traditional log records, and making the remote control process highly verifiable. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic flow chart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] Please refer to the attached Figure 1 , the embodiments of the present invention provide a remote control method for a mine trackless rubber-tyred vehicle based on 5G communication, including the following steps: S1. The vehicle completes identity authentication through the quantum key distribution protocol and generates a quantum key, and dynamically accesses the 5G-TSN network based on the authentication result. The 5G-TSN network is configured as a distributed communication framework supporting federated learning; Specifically, before starting the remote control process, the mine trackless rubber-tyred vehicle first establishes a quantum communication link with the security authentication center. The quantum communication link includes a laser emission unit, a polarization modulation module, and a single-photon detection module. The polarization modulation module is physically connected to the laser emission unit, outputs to a free space or an optical fiber channel, and is received by the single-photon detection module at the receiving end.

[0024] The vehicle and the security authentication center perform quantum key distribution based on the improved BB84 protocol. The improved BB84 protocol includes four polarization states: 0°, 90°, 45°, and 135°. Each polarization state represents a binary bit in two non-orthogonal bases. The vehicle end modulates and sends polarized photons in a random sequence, and the authentication center measures with a random basis to complete the generation of the original key.

[0025] In the original key comparison link, both parties exchange basis information through a public channel, and select the data bits with consistent measurement bases as candidate key bits. To prevent photon number splitting attacks, a decoy state mechanism is introduced, and signal photons (i.e., decoy states) with randomized intensities are inserted in part of the transmission process to measure the detection probability and judge potential attack behaviors.

[0026] Key collaboration is completed through information negotiation and error correction protocols, such as the Cascade or LDPC algorithm. The final key is used for data encryption, and the public key is issued by the certification center for subsequent integrity verification.

[0027] After identity authentication is completed and the key is successfully generated, the vehicle initiates a dynamic access request. The access process is based on the access control module in the 5G-TSN network, which includes an access gateway, an authentication and authorization component, and a time slot allocator.

[0028] The vehicle encrypts the communication request with the generated quantum key, and the authentication gateway decrypts it to verify its validity. After verification, according to the real-time channel quality parameters in the area where the vehicle is located, including signal-to-noise ratio (SNR), multipath fading value, and delay margin, the TSN communication time slot resources are dynamically allocated.

[0029] The TSN time slot priority is determined according to the following formula: ; Where: is the time slot priority score; is the distance between the vehicle and the nearest vehicle at present (unit: m); is the roadway curvature, and the value range is [0, 1], 0 represents a straight line, and 1 represents the minimum turning radius; is the current signal-to-noise ratio, unit is dB; are the corresponding weight coefficients respectively, satisfying: ; The weight coefficients are determined by an online optimization algorithm based on reinforcement learning. The reinforcement learning algorithm minimizes the objective function through an interactive iteration method: ; Where, represents the end-to-end communication delay, unit is ms; represents the packet loss rate, unit is % (percentage); is the control weight factor, satisfying ; By continuously adjusting the value, make reach the minimum to ensure the real-time performance and stability of network resource allocation.

[0030] The allocated TSN time slots are sent by the controller to the 5G communication module on the vehicle side. The module includes a receiving antenna, a beamforming controller, and a time slot synchronization clock. The time slot information is cross-mapped through the physical layer frame structure and the MAC layer synchronization mechanism to ensure that the communication module can complete data sending and receiving within the specified period.

[0031] The 5G-TSN network provides a data synchronization channel that supports federated learning for vehicles. Each vehicle conducts data interaction as an independent node, and all communication data uses symmetric encryption based on quantum keys to further enhance security.

[0032] All authentication requests, encryption parameters, and TSN scheduling information are coordinated through the control plane of the 5G core network. The data plane directly schedules resources and completes communication. The control plane and the data plane are logically isolated through a network slice manager to support the reliability requirements of multi-vehicle concurrent access.

[0033] S2. Based on quantum keys, vehicles collect roadway environment data in real time, encrypt it to generate encrypted data packets, and upload them to the edge server through the established 5G-TSN network. Specifically, first, the vehicle collects three-dimensional point cloud data of the roadway through a LiDAR (Light Detection and Ranging) sensor. The LiDAR sensor uses laser beams to scan the environment and generates a three-dimensional point cloud data set, accurately reflecting the roadway morphology and environmental characteristics. At the same time, the vehicle also obtains its pose data through a UWB positioning module. These data are used to determine the specific position, pose of the vehicle in the mine, and its relative relationship with the surrounding environment. The acceleration information of the vehicle is monitored in real time through an Inertial Measurement Unit (IMU) to obtain the vehicle's motion state.

[0034] Secondly, in the data encryption process, all collected data will be encrypted to ensure its security during transmission. The encryption uses encryption technology based on quantum keys, where the quantum key is used to encrypt the data to prevent the data from being intercepted or tampered with by a third party during transmission. Before encryption, the data integrity is verified through a public key. The specific verification method is to calculate the hash value of the data and compare it with the encrypted hash value to ensure that the data has not been lost or damaged during collection, storage, or transmission.

[0035] Once the data is encrypted, the vehicle uploads the encrypted data packet to the edge server through the 5G-TSN network. At this time, the upload process takes into account the network bandwidth and latency requirements. The 5G-TSN network has the characteristics of high bandwidth, low latency, and high reliability, which can ensure the rapid transmission of data. At the same time, when uploading data, based on the priority queue strategy of the 5G-TSN network, the system divides the priority according to the data type. Among them, the three-dimensional point cloud data is assigned the highest priority, the vehicle pose data is the second priority, and the acceleration data is the lowest priority.

[0036] The uploaded encrypted data packets use the following mathematical model for dynamic allocation of network resources to ensure that different data types can be processed in a timely manner in different priority queues: ; ; Among them, represents the real-time queue length of the point cloud data; represents the real-time queue length of the pose data; represents the bandwidth occupancy ratio of the point cloud data; represents the bandwidth occupancy ratio of the pose data.

[0037] S3. Based on the encrypted data packets uploaded in S2 and the edge server, construct a digital twin model to predict the dynamic changes of the environment. At the same time, generate a global collaborative control strategy through federated learning; Specifically, first, the edge server starts to construct a digital twin model by receiving encrypted data packets uploaded from multiple mine trackless rubber-tired vehicles. Specifically, the uploaded data includes three-dimensional point cloud data collected by LiDAR sensors, vehicle pose information, and acceleration information. The edge server performs voxel grid downsampling processing on these encrypted point cloud data to reduce the computational burden and extract key information, generating an accurate environmental model.

[0038] The mathematical formula for voxel grid downsampling processing is as follows: ; Among them, represents the representative point of each voxel after downsampling; represents the th point in the voxel grid, is the number of points within the voxel; Through this process, key environmental features can be effectively extracted.

[0039] Next, based on the extracted key feature point set, the edge server constructs a three-dimensional grid map and further predicts the deformation of the roadway. This prediction process is trained through a network model to achieve accurate simulation of the changes in the mine environment in the future for a period of time. The model training process combines the federated learning algorithm to achieve multi-vehicle collaborative training and global parameter aggregation.

[0040] Federated learning process In the collaborative computing process of the edge server, the federated learning technology is used to locally train the data of each mine vehicle. Each vehicle uses local sensor data for model training and encrypts and uploads the training results to avoid data privacy leakage. The server performs weighted averaging on the encrypted model parameters uploaded from each vehicle to generate a global updated model.

[0041] The mathematical formula for the federated learning process is as follows: ; Among them, is the update parameter for the global model, is the model parameter after local training of the th vehicle, is the weighting coefficient of vehicle ; The weight coefficient is calculated based on the Euclidean distance between the vehicle and the mine collapse risk area. Vehicles with a closer distance obtain a higher weight. The weight calculation formula is: ; Among them, is the Euclidean distance between the th vehicle and the collapse risk area. The edge server performs weighted averaging on the model parameters of each vehicle based on these weights to generate a globally optimized model.

[0042] Global collaborative control strategy generation Once the global model is updated, the edge server distributes the new global model parameters to each vehicle for iterative training until the model converges. During this process, each vehicle will continuously optimize using local data to ensure the effectiveness and adaptability of the model. Finally, through the training of the global model, a set of global collaborative control strategies are generated, which are used to optimize the motion control of the vehicle, including parameters such as steering and acceleration.

[0043] Specifically, the generation of the control strategy is based on the global model parameters and is carried out through the following formula: ; Among them, represents the parameters of the global collaborative control strategy; represents the control function of the th vehicle in the current environment; it is optimized based on the vehicle's state (such as position, speed, etc.); by minimizing the control errors of all vehicles, the optimal global collaborative control strategy is obtained.

[0044] S4. Execute the motion instruction based on the global collaborative control strategy and trigger the dynamic adjustment mechanism using the roadway environment data to update the vehicle steering and speed parameters; Specifically, the dynamic adjustment mechanism calculates the adjustment amounts of the vehicle steering angle and acceleration by monitoring real-time environment data such as vehicle spacing and roadway curvature and combining the global collaborative control strategy. The purpose of this mechanism is to optimize the smoothness and efficiency of vehicle driving based on parameters such as the relative positions between vehicles, the geometric characteristics of the roadway, and vehicle speed while ensuring safe driving. First, the dynamic adjustment mechanism uses the following mathematical model to calculate the adjustment amounts of the steering angle and acceleration: ; Among them, is the adjustment amount of the steering angle, in degrees (°); is the vehicle spacing, in meters (m); is the roadway curvature, in meters (m); is the adjustment amount of the acceleration, in meters per second squared (m / s ² ); is the current speed of the vehicle, in meters per second (m / s); This model first considers the relative positions of the vehicles , as well as the geometric shape of the roadway . Through this model, the steering angle can be dynamically adjusted to ensure that the driving direction of the vehicle matches the actual roadway environment and avoid the vehicle deviating from the safe track.

[0045] Secondly, for the adjustment of acceleration, the dynamic mechanism considers the vehicle spacing and the current speed . As the vehicle moves forward, if it is detected that the distance to the surrounding vehicles is too close or the roadway curvature changes greatly, the acceleration adjustment mechanism can optimize the acceleration value to ensure that the vehicle can drive smoothly and avoid dangers caused by too fast or too slow speed.

[0046] Control instruction execution After calculating the adjustment amounts of the steering angle and acceleration through the dynamic adjustment mechanism, the control system transmits these adjustment amounts to the underlying controller of the vehicle. The underlying controller uses a closed-loop control strategy to further optimize the actual executed steering angle and acceleration according to the input adjustment amounts (steering angle and acceleration adjustment amount).

[0047] Specifically, the execution process of the control instruction follows the following control law: ; ; Among them, is the actual steering angle of the vehicle, in degrees (°); is the previous steering angle of the vehicle, in degrees (°); is the adjustment amount of the steering angle, in degrees (°); is the actual acceleration of the vehicle, in meters per second squared (m / s ² ); is the previous acceleration of the vehicle, in meters per second squared (m / s ² ); is the adjustment amount of the acceleration, in meters per second squared (m / s ² ); This control law adjusts the behavior of the vehicle through real-time feedback to ensure that the vehicle can adapt and execute the optimal steering and speed regulation under different environmental conditions.

[0048] S5. Optimize the beamforming parameters and TSN time slot allocation of 5G communication according to the feedback data output by the dynamic adjustment mechanism, and at the same time encrypt and store the operation logs through the blockchain to form a closed-loop audit link; Specifically, first establish a beamforming optimization model, and calculate the beam direction angle adjustment amount based on the predicted boundary distance between the current position of the vehicle and the landslide risk area.

[0049] Let the current speed of the vehicle be (unit: m / s), the heading angle be (unit: rad), and the predicted distance from the current position of the vehicle to the boundary of the landslide risk area be (unit: m), then the beam direction angle adjustment amount is expressed as follows: ; Among them, represents the angle correction value that needs to be made in the current communication beam direction.

[0050] Subsequently, update the current beam direction angle , which is based on the previous step direction angle and the adjustment amount , and is calculated by the following formula: ; Among them, is the dynamic step coefficient; it is set according to the real-time channel state information (such as signal-to-noise ratio SNR), and its value range is , to control the sensitivity of the direction adjustment and ensure that there is no communication jitter or interruption during the beam adjustment process.

[0051] The above beamforming module is arranged in the beam controller of the vehicle communication terminal, and the controller includes a direction adjustment part and a channel feedback interface. The direction adjustment part works in coordination with the channel information obtained in the channel feedback interface based on the vehicle position and speed information output in step S4.

[0052] In terms of TSN time slot resource allocation, the system first sets a priority queue according to the sensor data type; the data types include: point cloud data (highest priority), vehicle pose data (secondary priority), and acceleration data (lowest priority).

[0053] According to the real-time queue length of each type of data in the network queue cache, the system dynamically calculates its bandwidth occupancy ratio adjustment parameter, and the bandwidth allocation ratio satisfies the following constraint conditions: ; Combined with the queue length feedback mechanism, the bandwidth ratio adjustment formula for point cloud and pose data is as follows: ; Among them, represents the dynamic bandwidth allocation ratio of point cloud data; represents the dynamic bandwidth allocation ratio of pose data; represents the real-time queue length of point cloud data at the current moment; represents the real-time queue length of pose data at the current moment; To prevent a tiny positive number in the denominator, the remaining bandwidth is allocated to acceleration data, that is: ; The TSN scheduling controller includes a time slot allocation unit and a priority queue configuration unit. The time slot allocation unit divides the transmission window period of the physical link according to the above bandwidth ratio; the priority queue configuration unit dynamically maps the data to the corresponding sending queue according to the sensor type.

[0054] To ensure the integrity and traceability during the communication optimization process, the system is also provided with a blockchain audit unit, which includes an operation record module and an encrypted storage module. The operation record module captures the vehicle control data and the communication parameter adjustment log in real time. The encrypted storage module calculates the hash value of the operation log based on SHA-256 and calls the distributed ledger interface to write it into the blockchain. Each adjustment operation is encapsulated as an immutable data block, forming a closed-loop audit link to ensure the security and controllability of the entire life cycle of the system.

[0055] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote control method for a mine trackless rubber-tyred vehicle based on 5G communication, characterized in that, It includes the following steps: S1. The vehicle completes identity authentication through the quantum key distribution protocol and generates quantum keys, and dynamically accesses the 5G-TSN network based on the authentication result. The 5G-TSN network is configured as a distributed communication framework supporting federated learning; S2. Based on the quantum keys, the vehicle collects roadway environment data in real time, encrypts it, generates encrypted data packets, and uploads them to the edge server through the established 5G-TSN network; S3. Based on the encrypted data packets uploaded in S2, the edge server constructs a digital twin model to predict the dynamic changes of the environment. At the same time, a global collaborative control strategy is generated through federated learning; S4. Execute the motion instruction based on the global collaborative control strategy, and trigger the dynamic adjustment mechanism using the roadway environment data to update the vehicle steering and speed parameters; S5. According to the feedback data output by the dynamic adjustment mechanism, optimize the beamforming parameters of 5G communication and the TSN time slot allocation, and at the same time encrypt and store the operation logs through the blockchain to form a closed-loop audit link.

2. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, wherein The execution of the quantum key distribution protocol in step S1 includes: Generate an encryption key through an improved BB84 protocol. Among them, the vehicle and the security authentication center exchange polarization-encoded quantum states, negotiate and generate a key after screening and matching bases, and resist photon number splitting attacks through decoy state detection; The security certification center distributes the public key to the vehicle for subsequent data integrity verification.

3. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, wherein, The dynamic access to the 5G-TSN network in step S1 includes: Encrypt communication requests based on the generated quantum key pairs, and dynamically allocate TSN time slots according to real-time signal quality parameters, with time slot priorities Calculated by the following formula: ； Among them, is the vehicle spacing, is the roadway curvature, is the signal-to-noise ratio; identifies the priority score or selection probability of the candidate path , and the larger the value, the higher the priority; the weight coefficient is optimized online through a reinforcement learning algorithm, and the optimization goal is to minimize the weighted sum of the end-to-end communication delay and the packet loss rate.

4. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, characterized in that, The roadway environment data in step S2 includes: 3D point cloud data collected by LiDAR, vehicle pose obtained by UWB positioning module , and acceleration information of the inertial navigation unit; Before data encryption, it is necessary to verify the data integrity through the public key The verification method is to calculate the data hash value and compare it with the encrypted hash value.

5. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, wherein, When uploading encrypted data to the 5G-TSN network in step S1: Divide the priority queue according to the data type. Among them, the point cloud data is assigned the highest priority, the pose data is assigned the second priority, and the acceleration data is assigned the lowest priority; The bandwidth allocation ratio of the priority queue is dynamically adjusted through the following formula: ； wherein, are respectively the real-time queue lengths of the point cloud and the pose data; are respectively the bandwidth occupancy ratios of the point cloud, the pose, and the acceleration data.

6. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, characterized in that The specific process of constructing the digital twin model in step S3 includes: Downsample the encrypted point cloud data uploaded by multiple vehicles through voxel grid and extract the key feature point set ; Based on the feature point set Construct a three-dimensional grid map And through The network predicts the future The roadway deformation parameters within ; The updated global model parameters are sent to each vehicle for iterative training until the model converges.

7. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, characterized in that, The federated learning process in step S3 includes: The edge server distributes the initial global model parameters to each vehicle. Each vehicle conducts local training based on the local roadway environment data and encrypts and uploads the model parameters to the edge server; The edge server calculates the aggregation weight dynamically according to the Euclidean distance between the vehicle and the landslide risk area and generates updated global model parameters through weighted average: ​​ ； Among them, is the weight of the global model in the th iteration, which represents the global model weight merged from the weights of each local model in the current iteration; is the th local model parameter of the th vehicle in the is the smoothing coefficient, is the total number of participating vehicles; The updated global model parameters are sent to each vehicle for iterative training until the model converges.

8. The remote control method for a mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, characterized in that, The dynamic adjustment mechanism in step S4 includes: Vehicle Spacing Based on Real-Time Monitoring and roadway curvature , the steering angle adjustment amount is calculated through the following formula and the acceleration adjustment amount : ； Wherein, is a preset safety distance, is a reference speed, is a weight coefficient determined by offline calibration, is the current real-time speed of the vehicle.

9. The remote control method for a mine trackless rubber-tired vehicle based on 5G communication according to claim 1, characterized in that, The vehicle steering and speed parameters in step S4 include the steering angle adjustment amount and the acceleration adjustment amount , and the specific steps include: Input and to the vehicle's low-level controller, and update the actual steering angle and acceleration : ； Among them, is the state variable in the th iteration, representing the new state updated based on the state of the previous time step at the current time step; is the state variable in th iteration, representing the state of the device at the previous time step; is the amount of change of the state variable in the th iteration, representing the new state update amount; is the auxiliary variable associated with the state variable , representing a certain state or input information in the th iteration; is the change amount of the variable in the state update, representing the increment from to during the iteration process of ; is the control period and satisfies .

10. The remote control method of the mine trackless rubber-tyred vehicle based on 5G communication according to claim 1, characterized in that, The steps for optimizing the beamforming parameters of 5G communication in step S5 include: Vehicle position output according to the dynamic adjustment mechanism From the predicted boundary of the landslide risk area Distance , calculate the adjustment amount of the beam direction angle , the formula is: ； Among them, is the current speed of the vehicle (unit: m / s); is the current heading angle of the vehicle; is the predicted distance from the vehicle position to the boundary of the landslide risk area (unit: m); Update the current beam direction angle based on the adjustment amount: ； Among them, represents the beam parameters in the th iteration; represents the beam parameters in the th iteration; represents the learning rate or adjustment factor, which is used to control the step size of each iteration; represents the change amount of the beam parameters in the current step size, which is dynamically set according to the channel state information to .

Images