A method for installing a coal mine underground roadway passive metasurface

Abstract

The application discloses a kind of coal mine underground roadway passive metasurface and its installation method, it is related to wireless communication technical field, comprising: collection roadway environment static data, mine car operation time series data, signal transmission time series data;Establish roadway time series three-dimensional model;Planning conventional main link and time series alternative link;In roadway middle deployment multiple passive metasurfaces, passive metasurface includes adjustable metasurface and fixed metasurface;From mine production management system obtains mine car real-time data, input mine car real-time data into time series scheduling model, generates scheduling time series table;According to scheduling time series table, in turn control each adjustable metasurface adjustment orientation angle, to switch between conventional main link and time series alternative link.The method of the application does not need to rely on fixed long path, and considers signal stability and scene adaptability.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to a method for wireless communication using passive metasurfaces in underground coal mines. Background Technology

[0002] Underground communication in coal mines is a core support for ensuring safe production, personnel scheduling, and equipment monitoring. Its stability directly affects the safety and efficiency of underground operations. Existing underground communication technologies mainly rely on devices such as repeaters, leaky cables, and passive electromagnetic metasurfaces to extend signal coverage. Among these, passive electromagnetic metasurfaces are widely used to solve problems such as roadway obstruction and signal attenuation due to their passive characteristics, flexible deployment, and low cost.

[0003] However, during the operation of underground rail transport vehicles, the electromagnetic metasurfaces on the signal link will be blocked one by one, resulting in the physical blockage of the fixed reflection path. This, combined with the electromagnetic interference generated by the vehicle's own motor and pantograph arc, as well as factors such as base station switching delay and multipath effect, can easily cause signal interruption, seriously affecting critical operations such as sensor data transmission and the issuance of unmanned driving control commands.

[0004] To address this issue, a fixed electromagnetic metasurface deployment method can be adopted, bypassing the mining truck's path. This involves considering scenarios such as different mining truck sizes and multiple trucks traveling in parallel during the design phase, planning fixed reflective links that do not pass directly above or in front of the mining truck's track, attempting to achieve coverage for all mining truck occlusion scenarios through a one-time deployment of passive metasurfaces. However, this deployment method has significant drawbacks: to adapt to the worst-case scenario of "largest mining truck size + multiple trucks traveling in parallel + extreme occlusion," the fixed link must adopt a design of "large diffraction angle + long transmission path," resulting in extremely poor signal quality in normal scenarios, i.e., when no mining trucks are passing, only small mining trucks are passing, or mining trucks have not entered the occlusion area. Specifically, the transmission distance of a fixed bypass link is usually 30%-60% longer than that of a straight path. The attenuation coefficient of electromagnetic waves in the 2.4G / 5.8G band underground is about 0.2-0.3dB / m. An excessively long transmission path will increase the total signal attenuation by 5-18dB. At the same time, the number of reflections will exacerbate signal loss and multipath interference, causing the signal strength to be close to or even lower than the communication threshold (-80dBm) in normal scenarios. This results in frequent signal stuttering, weak signals, and other problems, which cannot meet the communication stability requirements of critical underground operations. Summary of the Invention

[0005] This application provides a method for installing a passive metasurface in underground coal mine roadways, which solves the problem of poor signal quality in the existing electromagnetic metasurface deployment method that uses a fixed method and bypasses the mine car's travel path.

[0006] This application provides a method for installing a passive metasurface in underground coal mine roadways, including:

[0007] Collect static environmental data of underground coal mine roadways, timing data of mine car operation, and timing data of signal transmission in roadways at different time periods;

[0008] A three-dimensional mesh model of the tunnel is established based on static environmental data of the tunnel. The timing data of mine car operation and signal transmission are imported into the three-dimensional mesh model of the tunnel to form a time-series three-dimensional model.

[0009] Based on the timing patterns of mine car occlusion and signal transmission, conventional main links and temporal alternative links are planned in the temporalized 3D model. Conventional main links are used during periods without occlusion, and temporal alternative links are used during periods of mine car occlusion.

[0010] Multiple passive metasurfaces are deployed in the tunnel based on the conventional main link and the time-sequential alternative links. The passive metasurfaces include adjustable metasurfaces and fixed metasurfaces. The adjustable metasurfaces can adjust the orientation angle, while the fixed metasurfaces have a fixed orientation angle.

[0011] Real-time data of mining trucks is obtained from the mine production management system. The real-time data of mining trucks is input into the time-series scheduling model to predict the key time points of shading of each adjustable metasurface and generate a scheduling time series table according to the direction of mining truck travel.

[0012] According to the scheduling sequence, each adjustable metasurface is controlled to adjust its orientation angle in sequence to switch between the regular main link and the time-sequential alternative link.

[0013] The method for installing a passive metasurface in underground coal mine roadways disclosed in this application has the following advantages:

[0014] 1. Balanced signal quality across all scenarios: In normal scenarios, the adjustable metasurface maintains its initial angle to achieve straight-line reflection, avoiding long-distance transmission and multiple reflection losses caused by fixed bypass links. The signal strength is stable at around -70dBm, far from the communication threshold, completely solving the defects of signal lag and weak signal in normal scenarios with fixed links. In extreme scenarios such as mine truck obstruction, the link is switched and the angle is adjusted to bypass the obstruction through time-sequential scheduling, without relying on fixed long paths, thus balancing signal stability and scenario adaptability.

[0015] 2. Dynamic Adaptation Without Compromise: There is no need to sacrifice conventional performance to cover extreme scenarios. By predicting the driving data of mining trucks and dynamically adjusting the metasurface angle, it can flexibly adapt to different mining truck sizes, multiple trucks running in parallel, and temporary changes in the state of mining trucks. At the same time, it can cope with fluctuations in the tunnel environment and electromagnetic interference, avoiding the limitations of fixed links where "static design cannot adapt to dynamic changes".

[0016] 3. Efficient resource reuse: Through the combination of "tunable metasurface + fixed metasurface", the fixed metasurface can adapt to multiple links without the need to deploy dedicated metasurfaces separately for each scenario. This avoids the redundant costs of adding equipment after the fixed link scenario changes, and achieves efficient resource utilization throughout the entire life cycle. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, 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.

[0018] Figure 1 This is a flowchart illustrating a method for installing a passive metasurface in an underground coal mine roadway, as provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] Figure 1 This application provides a flowchart of a method for installing a passive metasurface in an underground coal mine roadway, comprising the following steps:

[0021] S100 collects static environmental data of underground mine roadways, timing data of mine car operation, and timing data of signal transmission in roadways at different time periods.

[0022] For example, the static environmental data of the tunnel includes three-dimensional point cloud data of the tunnel, key cross-sectional dimensions, tunnel wall material, obstacle distribution, location and curvature of bends / intersections, and coordinates of power supply convenience points. After collecting the static environmental data of the tunnel, data processing is performed, including point cloud denoising, registration and stitching, segmentation simplification, and static data standardization.

[0023] Specifically, the detailed data processing flow is as follows.

[0024] Point cloud denoising employs a statistical filtering algorithm combined with radius filtering to ensure point cloud purity. The statistical filtering algorithm calculates the number of neighboring points within a 0.1m radius for each point and removes isolated points with a mean ± 3 standard deviations. The radius filtering algorithm, on the other hand, removes dense noise points with fewer than 5 neighboring points within a 0.05m radius.

[0025] Registration and stitching involves extracting key feature points from the point cloud of each station using the SIFT (Scale Invariant Feature Transform) algorithm, such as road corners and support connections. The RANSAC (Random Sample Consensus) algorithm is then used to remove incorrect matching pairs, and the ICP (Iterative Closest Point) algorithm is used to minimize the Euclidean distance error of the point cloud. The stitching error is required to be ≤ ±5cm to form a complete road point cloud model.

[0026] Segmentation simplification is based on a region growing algorithm. It segments the point cloud into regions such as lane walls, top, bottom, and obstacles according to geometric features, such as normal vectors and curvature. Then, a grid simplification method (such as uniform sampling) is used to reduce the number of point clouds. The amount of data after simplification is reduced by 60%-70%, while maintaining a geometric feature similarity of ≥95%.

[0027] Static data standardization involves extracting the spatial coordinates, dimensional parameters, and three-dimensional coordinates of the power supply point of the tunnel, and converting them into standardized data in a three-dimensional coordinate system for the tunnel. This coordinate system has the tunnel entrance as the origin and the track extension direction as the X-axis, and is finally stored in PLY (polygon file) format.

[0028] The mining truck runtime sequence data comes from the mine production management system's historical operation data for the past 3 months, including the mining truck ID, timestamp, real-time location, driving speed, driving direction, operating status (e.g., normal / deceleration / acceleration / stop), and mining truck specifications (including length, width, and height).

[0029] After acquiring the timing data of the mining truck operation, the timing data processing is performed according to the following procedure:

[0030] Data cleaning. A sliding window filter with a window size of 5 seconds is used to remove outlier data, such as position jumps >10m, speeds >30km / h, and contradictory state logic, while retaining continuous and stable time series.

[0031] Temporal feature extraction includes basic temporal features and occlusion temporal features. The basic temporal features are the average travel speed of the mining truck (10-20 km / h) and the acceleration range (≤5 km / h) of the mining truck on different road sections. 2 The fixed route's travel cycle (e.g., passing through a metasurface area once per hour). The occlusion timing features are based on the coordinates of the mine car's position and the candidate deployment area of ​​the metasurface. The arrival time of the mine car in the occlusion area, the duration of occlusion, and the departure time from the occlusion area are extracted to form mine car-metasurface occlusion timing pairs. For example, if a mine car travels at 10 km / h, is 15 m long, the metasurface is deployed at 300 m, and the occlusion area ranges from 290 m to 310 m, then the occlusion arrival time = (290 - current position) / speed, and the occlusion duration = 15 / (speed × 1000 / 3600).

[0032] Time-series dataset construction. A time-series dataset of mining truck runtime is constructed in the format of timestamp-miner ID-location X-velocity-direction-occlusion metasurface ID-occlusion arrival time-occlusion duration, containing 100,000 valid samples.

[0033] Signal transmission timing data is obtained by collecting signal strength (RSRP), signal-to-noise ratio (SNR), and packet loss rate at different time periods (no mine cars, mine cars passing by, and periods of high electromagnetic interference) using a mine signal tester (2.4G / 5.8G dual-band, sampling frequency 10Hz). These data record the changing patterns of signal parameters over time.

[0034] After acquiring the signal transmission timing data, the analysis is performed according to the following timing sequence:

[0035] During normal periods, i.e. when there are no mining trucks: Statistically measure the stable values ​​of signal parameters, such as RSRP=-70dBm±2dBm and SNR=22dB±1dB, to determine the signal baseline values ​​under normal scenarios.

[0036] Mining car passage time: Analyze the change curves of signal parameters with the timing of mining car obstruction. For example, when the mining car approaches the metasurface, RSRP decays at a rate of 0.5dBm / s, drops to below -95dBm when obstructed, and recovers to the reference value within 1 second after leaving.

[0037] Interference period: Record the occurrence time and signal parameter fluctuation amplitude of interference events such as pantograph arcing and inverter start-up and shutdown. For example, if the SNR drops from 22dB to 10dB during interference, an interference timing feature library is formed.

[0038] S110: Based on static data of the tunnel environment, a three-dimensional mesh model of the tunnel is established. The timing data of the mine car operation and the timing data of signal transmission are imported into the three-dimensional mesh model of the tunnel to form a timing-series three-dimensional model.

[0039] For example, after collecting static data of the tunnel environment, timing data of mine car operation, and timing data of signal transmission, coordinate unification and time synchronization processing are performed to form a timing-related dataset. Based on the timing-related dataset, a three-dimensional mesh model of the tunnel and a timing-series three-dimensional model are established.

[0040] Specifically, coordinate unification unifies the coordinate system of static roadway data, mine car time-series data, and signal time-series data into a three-dimensional roadway coordinate system, ensuring accurate spatial correspondence. Time synchronization, on the other hand, adopts the NTP (Network Time Protocol) time synchronization protocol to calibrate the timestamp accuracy of all data to the millisecond level, requiring a synchronization error of <10ms, forming a time-series correlation dataset of static environment-dynamic mine car-signal changes, providing a unified data foundation for subsequent time-series modeling.

[0041] Based on the static tunnel data in the time-series correlation dataset, a 1:1 three-dimensional mesh model of the tunnel was generated using the Poisson reconstruction algorithm, restoring the tunnel's length, width, height, and inner wall contour, with a dimensional deviation ≤ ±5cm. Then, electromagnetic parameters were assigned to different regions of the model; for example, the concrete tunnel wall had a dielectric constant of 4.5, a permeability of 1, and an attenuation coefficient of 0.25dB / m; the metal support had a dielectric constant of ∞, a permeability of 1, and a reflection coefficient of 0.95, thus recreating the actual electromagnetic propagation environment. Next, power supply convenience points, obstacles, base station locations, and receiver locations were mapped one by one into the model, clarifying the three-dimensional coordinates and spatial relationships of each element.

[0042] After the above processing, the mining truck runtime sequence dataset is imported into the 3D mesh model of the tunnel. The model reconstructs the mining truck's trajectory and occlusion timing, dynamically displaying the occlusion status of each metasurface candidate region at different time points according to the logic of occlusion arrival time - occlusion duration - departure time: unoccluded / occluded / departed. Next, based on signal transmission timing data, the changes in signal parameters over time are superimposed on the model. For example, when the mining truck occludes the metasurface, the signal strength dynamically attenuates from the baseline value and dynamically recovers after departure, intuitively presenting the correlation between the mining truck timing and the signal timing. Finally, interference events from the interference timing feature library (such as the high incidence of pantograph arcing from 14:00 to 14:05) are incorporated into the model to simulate signal fluctuations during interference periods.

[0043] After establishing the above temporal 3D model, the model needs to be temporally verified and optimized according to the following process:

[0044] Timing accuracy verification. Compare the occlusion timing of the mine car in the model with the actual occlusion timing measured on site to ensure that the occlusion arrival time error is ≤ ±0.2 seconds and the occlusion duration error is ≤ ±0.1 seconds.

[0045] Signal timing verification. The model simulates a mine car passing over a metasurface, and the signal intensity variation curve is compared with the field-measured curve to ensure that the errors in attenuation rate, minimum value, and recovery time are all ≤10%.

[0046] Lightweight model optimization. Redundant details unrelated to timing and scheduling, such as tunnel wall textures, are removed, while the core structure and timing scene are preserved to ensure model simulation efficiency, requiring a single timing simulation time of less than 10 seconds.

[0047] S120, based on the timing patterns of mine car occlusion and signal transmission, plans a conventional main link and a timing-sequential alternative link in a timing-sequential 3D model. The conventional main link is used during periods without occlusion, and the timing-sequential alternative link is used during periods of mine car occlusion.

[0048] For example, the planning principles for the conventional main link are to minimize signal attenuation and the number of reflections. The planning principles for the time-sequential alternative links include the timing adaptation principle, the redundancy coverage principle, and the signal compliance principle. The timing adaptation principle is that each alternative link in the time-sequential alternative links corresponds to a set of mine car-metasurface occlusion timing pairs, ensuring that when a mine car occludes a metasurface in the link, the corresponding alternative link can be activated before the occlusion arrives. The redundancy coverage principle is to plan at least one alternative link for each critical metasurface of the conventional main link, while the alternative link has no risk of overlapping occlusion with the conventional main link. The signal compliance principle is that the transmission distance of the alternative link is less than or equal to 30% longer than that of the conventional main link, and the increase in signal attenuation is less than or equal to 5dB.

[0049] Specifically, the planning principle of conventional main links is to minimize signal attenuation, minimize the number of reflections, and ensure timing stability. The propagation path is selected with straight paths as the main path and a small number of reflections as the auxiliary path, to ensure that the signal parameters are stable at the reference value under normal scenarios, requiring RSRP≥-75dBm and SNR≥20dB.

[0050] The specific planning process can be represented as follows: Starting from the base station (0m), plan the main link along the track extension direction (X-axis), with ≤2 reflections and a transmission distance consistent with the tunnel length (1000m). For example, the main link is "base station → fixed metasurface PS1 at 150m → fixed metasurface PS2 at 450m → receiver (1000m)", and the signal strength is stable at around -69dBm during normal periods.

[0051] The specific planning process for time-series alternative link planning can be represented as follows: Based on the time-series 3D model, candidate deployment locations for metasurfaces are selected. For example, adjustable metasurfaces RS1, RS2, and RS3 are deployed at 300m, 600m, and 900m respectively, and fixed metasurfaces PS3 and PS4 are deployed at 750m and 850m respectively. For the main link metasurface PS1, alternative link 1 is planned: "Base station → RS1 → PS3 → Receiver". The metasurfaces RS1 and PS3 of this link will not be simultaneously blocked by the mining truck. For the main link metasurface PS2, alternative link 2 is planned: "Base station → PS3 → RS2 → Receiver". Similarly, it is ensured that there is no overlapping blockage in time sequence. The corresponding triggering time sequence condition is marked for each alternative link. For example, the triggering condition for alternative link 1 is "2 seconds before the mining truck arrives at the PS1 blocking area". After triggering, the link is activated and the angle of RS1 is adjusted.

[0052] After planning multiple alternative links, the link timing switch needs to be performed according to the following rules:

[0053] Switching trigger mechanism. Based on the minecart occlusion timing, an advance warning time is set, such as 1-2 seconds. When the remaining time for the minecart to reach the occlusion area of ​​the metasurface is less than or equal to the warning time, a link switch is triggered. For example, if the minecart arrives at the PS1 occlusion area at 14:00:02 and the warning time is set to 2 seconds, then alternative link 1 will be activated at 14:00:00.

[0054] Switching priority rules. Prioritize signal quality > timing compatibility > link length, meaning that the best alternative link with the best signal quality, the best timing compatibility with the mining truck, and the shortest transmission distance will be switched to first.

[0055] Switching latency requirements. The link switching process is optimized through model simulation to ensure that the total latency from triggering the switch to the stable operation of the backup link is <50ms, which is far below the retransmission threshold of the communication protocol, such as 200ms.

[0056] After obtaining the conventional main link and time-sequential alternative links, the following simulation verification is required:

[0057] Simulation scenario coverage. Three typical time-series scenarios are simulated: ① Normal period, no mining trucks, main link operates stably; ② Single mining truck operation, switching to backup links 1 and 2 according to the obstruction sequence; ③ Multiple trucks in parallel operation + interference, two trucks in parallel obstructing the main link, superimposed electromagnetic interference, switching to redundant backup links. Simulation metrics include recording the signal strength, SNR, switching delay, and number of interruptions for each link, ensuring: Under normal scenario, the main link signal strength ≥ -75dBm, fluctuation range ≤ ±1.5dB; Under obstruction scenario, the signal strength after activation of the backup link ≥ -80dBm, with no interruptions; Under interference scenario, the backup link's anti-interference capability meets the standards, i.e., SNR ≥ 15dB, packet loss rate ≤ 1%. For backup links with substandard signals in the simulation, the candidate deployment position or reflection angle of the metasurface is adjusted until the links meet the requirements in all time-series scenarios. A time-seriesd main / backup link planning diagram is output, clearly defining the link composition, switching timing conditions, and signal parameter thresholds.

[0058] S130: Multiple passive metasurfaces are deployed in the tunnel according to the conventional main link and the time-sequential alternative link. The passive metasurfaces include adjustable metasurfaces and fixed metasurfaces. The adjustable metasurfaces can adjust the orientation angle, while the fixed metasurfaces have a fixed orientation angle.

[0059] For example, before deploying the passive metasurface, deployment locations are first selected in the tunnel. The selection criteria include timing compatibility, power supply feasibility, and link compatibility. Then, the initial angle of the passive metasurface is calculated, and finally, the passive metasurface is deployed according to the deployment location and the initial angle.

[0060] Specifically, in the selection principle for deployment locations, timing adaptability means that the deployment location of the metasurface must match the timing of the mining truck's occlusion, ensuring that the mining truck occludes the metasurface one by one in the order of travel, rather than occluding multiple metasurfaces simultaneously, thus providing a time window for timing switching. For example, if deployed along the X-axis in the order of 150m (PS1), 300m (RS1), 450m (PS2), 600m (RS2), 750m (PS3), and 900m (RS3), when the mining truck travels at 10km / h, the occlusion interval is (300-150) / (10×1000 / 3600)=54 seconds, which is sufficient to complete the link switching.

[0061] Power supply feasibility is ensured by deploying adjustable metasurfaces only at locations with convenient power supply, namely 300m, 600m, and 900m, while fixed metasurfaces are deployed in non-powered areas, reducing power supply costs. Link adaptability is ensured by precisely matching the reflection points of the primary and backup links at the metasurface deployment locations, guaranteeing that the primary link signal is transmitted along the planned path and that the backup link can quickly form an effective reflection after activation.

[0062] The final deployment plan can be represented as follows:

[0063] Adjustable metasurface RS: 3 sets, deployed at power supply convenience points of 300m (RS1), 600m (RS2), and 900m (RS3) respectively, responsible for angle adjustment and link connection during timing switching.

[0064] Fixed metasurfaces PS: 6 sets, deployed at 150m (PS1), 450m (PS2), 500m (PS4), 750m (PS3), 850m (PS5), and 950m (PS6) respectively, to meet the reflection requirements of multiple primary and backup links.

[0065] Deployment sequence: According to the direction of the mine car's travel (positive X-axis direction), the metasurfaces are sorted as PS1→RS1→PS2→PS4→RS2→PS3→PS5→RS3→PS6, which clarifies the order of subsequent time-series scheduling.

[0066] The principle for calculating the initial angle of a metasurface is that the initial angle must simultaneously meet the requirements of optimal main link signal in normal scenarios and rapid switching of alternative links in occlusion scenarios. That is, the main link signal is stable during normal periods, and only minor adjustments to the angle are needed to switch to the alternative link when occlusion occurs.

[0067] The initial angles include the main link angle and the alternative link angles. The main link angle is based on the reflection path of the main link, calculated according to the specular reflection law (angle of incidence = angle of reflection), and combined with the coordinates of the base station, metasurface, and receiver in the 3D model. The initial elevation and azimuth angles of each metasurface are calculated. For example, the initial angles for PS1 (at 150m) are: elevation 25°, azimuth 0°, ensuring the main link signal strength ≥ -75dBm from base station → PS1 → PS2 → receiver. The alternative link angles are calculated for each alternative link, determining the target angle of the adjustable metasurface (the angle after switching) and storing it in the system. For example, the target angle for alternative link 1 of RS1 is: elevation 38°, azimuth 0°, ensuring the signal strength ≥ -80dBm from base station → RS1 → PS3 → receiver. Finally, the initial and target angles are bound to the mine car occlusion timing, forming a table of association between metasurface ID, initial angle, target angle, and trigger switching time, providing angle parameter basis for subsequent timing scheduling.

[0068] In the planned links, each metasurface is ordered according to its deployment sequence, such as PS1→RS1→PS2→…→PS6. Scheduling priorities are set, and as the mining truck travels, metasurface adjustments and link switching are triggered in priority order to avoid scheduling conflicts. For example, if the mining truck approaches PS1 first, the alternative link 1 corresponding to PS1 is scheduled first; after leaving PS1, the alternative link corresponding to RS1 is scheduled next.

[0069] The link needs to be switched according to the following coordination rules:

[0070] Overlapping time period handling. When the occlusion timing of adjacent metasurfaces overlaps by 0.5 seconds, if the mining truck moves away from PS1 and approaches RS1 at the same time, a collaborative logic of activation followed by shutdown is adopted. That is, the alternative link of RS1 is activated in advance, and the alternative link of PS1 is shut down after the mining truck has completely left PS1, ensuring seamless link connection.

[0071] Multi-metasurface collaboration. When mining trucks simultaneously block multiple passive metasurfaces, such as two trucks blocking PS1 and PS2 in parallel, multiple alternative links are activated, such as RS1→PS3→RS2. Through the coordinated angle of the adjustable metasurfaces, a relay reflection link is formed.

[0072] During the process of forming the metasurface deployment scheme according to the above link, it is necessary to mark the metasurface installation positions according to the deployment scheme, simulate the movement of mining trucks, and measure the signal strength and occlusion timing at different time points to ensure that the metasurface deployment positions match the occlusion timing of the mining trucks and that there is no risk of overlapping occlusion. Furthermore, the deployment scheme is imported into the time-series 3D model, simulating the movement of mining trucks at normal speeds to verify the timing logic of link switching, the accuracy of metasurface angle adjustment, and the stability of signal quality, ensuring that: the switching trigger time error is ≤ ±0.1 seconds; the metasurface angle adjustment accuracy is ≤ ±0.1°; and the signal strength during occlusion periods is ≥ -80dBm without interruption.

[0073] The final output is a time-sequential electromagnetic metasurface deployment scheme, which includes complete information such as metasurface deployment coordinates, type, initial angle, target angle, scheduling priority, and link switching timing logic.

[0074] According to the coordinates and initial angle of the deployment plan, fix the adjustable metasurface and the fixed metasurface, ensuring that the front of the metasurface faces the signal incident direction and the flatness error of the mounting surface is <1mm; the electric adjustment device (specifically a stepper motor) and control module of the adjustable metasurface are fixed on the back, and the wiring uses shielded cables to avoid electromagnetic interference.

[0075] After deploying the metasurfaces, a real-time channel feedback sensor is deployed next to each group of metasurfaces. A signal tester is deployed at the receiving end, and a position sensor is installed on the mining truck. This sensor synchronizes data with the production management system, and the timestamps of all sensors are synchronized to the millisecond level.

[0076] The underground edge gateway is deployed at 500m and its function is to aggregate the metasurface status, sensor data and mine car time series data. The ground server deploys the time series scheduling model and monitoring system to realize global time series management.

[0077] Furthermore, the structure of the passive metasurface can be referenced to the electromagnetic metasurface array in CN119401131A. The structure of the passive metasurface in this application is the same as that of the electromagnetic metasurface array in that patent. The difference is that all electromagnetic metasurface units contained in the electromagnetic metasurface in this application use the exact same unit parameters and arrangement.

[0078] S140: Obtain real-time data of mining trucks from the mine production management system, input the real-time data of mining trucks into the time-series scheduling model, predict the key time points of shading of each adjustable metasurface, and generate a scheduling time sequence table according to the direction of travel of the mining trucks.

[0079] For example, mine car data is extracted in real time at a frequency of 1Hz, including: mine car ID (unique identifier), timestamp (accurate to milliseconds), real-time location (X / Y coordinates, based on the three-dimensional model coordinate system of the tunnel), driving speed (km / h), driving direction (e.g., along the positive X-axis), operating status (normal driving / deceleration / acceleration / stopping), and scheduling route (e.g., main tunnel 1000m → main tunnel 2000m).

[0080] A sliding window filter is used for real-time data of mining trucks, with a window size of 5 seconds, to remove abnormal data, such as sudden changes in position exceeding 10m or speed exceeding 30km / h, while retaining continuous and stable mining truck operation data.

[0081] Next, the data will undergo the following standardization process:

[0082] Coordinate transformation. The mine car position data is converted from the global coordinate system of the production management system to the local coordinate system of the 3D model of the tunnel. This coordinate system takes the tunnel starting point as the origin and the track extension direction as the X-axis to ensure consistency with the coordinates deployed on the supersurface.

[0083] Feature engineering. Construct model input features, including: the distance between the current position and each metasurface (e.g., the distance D1 from the minecart to RS1, the distance D2 from RS2), the first derivative of the travel speed (acceleration a), the angle (θ) between the travel direction and the deployment direction of the metasurface, and the length of the minecart (L).

[0084] All data is consistent with the three-dimensional coordinate system and time-series correlation dataset format of the tunnel built in the early deployment phase to ensure compatibility with the input of the time-series scheduling model.

[0085] Furthermore, the time-series scheduling model is built on LSTM (Long Short-Term Memory) network + attention mechanism and has been trained on the training dataset.

[0086] Specifically, the input layer in the time-series scheduling model is used to receive the standardized mine car data features, which have a total of 8 input dimensions: distance Dn from the mine car to RSn, mine car speed v, acceleration a, angle θ of the driving direction, mine car length L, X coordinate of RSn, Y coordinate of RSn, and current channel SNR.

[0087] The hidden layer consists of two LSTM layers and one attention layer. LSTM layer 1 has 64 LSTM units, an activation function of tanh, and a dropout rate of 0.2, used to extract temporal features of the mining truck's movement, such as speed change trends and position movement patterns. LSTM layer 2 has 32 LSTM units, an activation function of tanh, and a dropout rate of 0.2, further exploring the correlation between temporal features and metasurface occlusion.

[0088] The attention layer adopts a multi-head attention mechanism with 4 attention heads and the weight parameters are initialized to a uniform distribution to highlight key features, such as the impact of mine car distance and speed on occlusion time.

[0089] The output layer is a fully connected layer with 32 neurons and a linear activation function. It outputs prediction results in four dimensions: occlusion warning time T_warning, occlusion start time T_start, occlusion end time T_end, and link recovery time T_recovery.

[0090] The model's other parameters were set as follows: learning rate of 0.001, batch size of 32, number of training epochs of 100, and early stopping patience of 10.

[0091] The training dataset was collected from three months of historical underground data, including mine truck movement data, metasurface occlusion records, and channel data, totaling 100,000 samples. Each sample contains 30 seconds of time-series data and corresponding key occlusion time point labels. The dataset was divided into a training set (70,000 samples), a validation set (20,000 samples), and a test set (10,000 samples) in a 7:2:1 ratio to ensure even distribution. Based on the mine truck positions and metasurface occlusion records in the historical data, the T-warning, T-start, T-end, and T-recovery values ​​for each RSn in each sample were manually labeled. For example, if a mine truck arrives at the starting point of the RS1 occlusion zone at 14:00:00, the label would be T-start = 14:00:00.

[0092] The training process of the model is as follows:

[0093] Loss function design: A hybrid loss function is adopted to balance the accuracy of time prediction and the success rate of link handover. The hybrid loss function includes: time prediction loss and link handover loss. The time prediction loss adopts MAE (mean absolute error), which is the average of the sum of the differences between the measured values ​​and the predicted values ​​at T warning, T start, T end and T recovery. The link handover loss adopts cross entropy. The hybrid inspection function is the weighted sum of the two loss functions.

[0094] Optimizer selection: The Adam optimizer is used with momentum parameters β1=0.9, β2=0.999, and weight decay coefficient λ=0.0001 to avoid gradient explosion during training.

[0095] The training steps are as follows:

[0096] Initialize the model parameters to follow a normal distribution N(0,0.01);

[0097] The training set is input into the model in batches of the specified size, and forward propagation is used to calculate the predicted output and the loss function.

[0098] Backpropagation is used to calculate the gradient, and the model parameters are updated using the Adam optimizer.

[0099] After each training epoch, the model performance is evaluated using a validation set, including time prediction error and switching success rate.

[0100] If the validation set loss does not decrease for 10 consecutive rounds, the early stopping mechanism is triggered to save the current optimal model.

[0101] Test the optimal model using a test set to ensure that the time prediction error is less than ±0.2 seconds and the link switching success rate is greater than or equal to 99%.

[0102] Furthermore, the timing scheduling model is deployed in the downhole edge gateway and TensorRT (Tensor Runtime) is used to accelerate the model.

[0103] Specifically, the trained model is embedded into the underground edge gateway, and TensorRT is used to accelerate the model, ensuring inference latency is <50ms. The model loads the metasurface initial angle-target angle correlation table and occlusion temporal feature library pre-computed during the early deployment phase to ensure that the prediction logic is consistent with the early deployment plan. Every 10,000 new samples are accumulated and added to the training set to retrain the model, for example, by fine-tuning the learning rate to 0.0001, the number of training epochs to 20, optimizing the model parameters, and adapting to changes in the mining truck's operating status.

[0104] Furthermore, after predicting the critical time points of obstruction, the critical time points of obstruction are corrected by combining real-time channel feedback data.

[0105] Specifically, the preprocessed real-time data of the mining trucks (30-second time series) is input into the timing scheduling model of the edge gateway. The model outputs T_warning, T_start, T_end, and T_recovery for each adjustable metasurface (RS1, RS2…RSn). T_warning is the trigger time for initiating metasurface adjustment, 1-2 seconds earlier than T_start; T_start is the time when the mining truck enters the metasurface obstruction zone and the main link begins to fail; T_end is the time when the mining truck leaves the metasurface obstruction zone and the main link signal recovers; T_recovery is the time when the backup link is closed and the connection is switched back to the main link. Finally, combined with real-time channel feedback data, if the current SNR < 15dB, it indicates signal interference, and the critical time points are corrected, such as advancing the T_warning by 0.5 seconds. If the mining truck's operating status changes, such as deceleration or stopping, the data is re-inputted into the model for inference, and the critical time points are updated.

[0106] Based on the direction of the mine car's travel (e.g., along the positive X-axis), the key time points of each adjustable metasurface are sorted in order of T-warning sequence to form a scheduling timetable, ensuring that scheduling instructions are executed in the order of mine car occlusion. Each record in the table includes metasurface number, T-warning, T-start, T-end, T-recovery, alternative link, and angle adjustment parameters, as shown in Table 1.

[0107] Table 1 shows an example of data from the scheduling sequence table.

[0108]

[0109] The scheduling time table is updated every 500ms during the scheduling process. Combined with the real-time position and speed changes of the mining trucks, the time points and adjustment parameters are dynamically adjusted to ensure scheduling accuracy.

[0110] S150 controls each adjustable metasurface to adjust its orientation angle according to the scheduling sequence table in order to switch between the regular main link and the time-sequential alternative link.

[0111] For example, the signal scheduling system, according to the scheduling sequence table, follows the progress of the mine car and sequentially performs early warning activation → angle adjustment → link switching → return to sleep operation on each adjustable metasurface to achieve seamless link relay.

[0112] The timing of the previously deployed metasurfaces is ordered, for example, PS1→RS1→PS2→RS2… to trigger scheduling, ensuring that the link switching order is completely matched with the previously planned occlusion timing.

[0113] Specific examples are as follows:

[0114] Phase 1: RS1 scheduling is executed, and the mining truck is about to block RS1.

[0115] T-warning (14:00:00). Activate backup link: The system issues a command to start the backup link from base station → PS3 → RS1 → receiver, activating the control module and real-time channel feedback sensor of RS1. The fixed metasurface PS3 does not require adjustment. Angle adjustment: The control module of RS1 drives the stepper motor to adjust the reflection angle according to the target angle (i.e., 38° / 0°) in the scheduling timetable. During the adjustment process, the angle sensor provides real-time feedback on the current angle. Adjustment stops when the deviation from the target angle is < ±0.1°. Link detection: The real-time channel feedback sensor collects the signal parameters of the backup link. If 10 consecutive sampling points meet RSRP ≥ -80dBm, SNR ≥ 15dB, and packet loss rate ≤ 1%, the backup link is confirmed to be stable. Pre-handover preparation: The signal reception priority of the receiver is switched to the backup link, while the main link remains operational to avoid handover interruption.

[0116] T begins (14:00:02). The mining truck enters the obstruction zone of RS1, and the main link is blocked by the metal body of the mining truck, causing the signal strength to drop sharply to <-95dBm. The system immediately issues a command to shut down the main link, and the receiving end completely switches to the backup link, where the signal strength stabilizes at -72dBm without interruption.

[0117] T ends (14:00:07). The mine truck leaves the RS1 obstruction area, the main link signal begins to recover, and the real-time channel feedback sensor detects that the main link RSRP ≥ -75dBm and SNR ≥ 18dB.

[0118] T resumes (14:00:07.5). The system issues a command, the receiving end switches back to the primary link and closes the backup link. The electric adjustment device of RS1 drives the angle back to the initial value (i.e., 25° / 0°), the control module enters sleep mode, and the scheduling loop of RS1 is completed.

[0119] Phase 2: RS2 scheduling execution (the mining truck is about to block RS2).

[0120] Warning T (14:00:05). Activate backup link: Start the backup link from base station → RS2 → PS5 → receiver, activating the control module and sensor of RS2. Angle adjustment: RS2 adjusts the reflection angle according to the target angle (i.e., 42° / 0°), adjustment time < 500ms, link detection confirms stability. Pre-handover preparation: The receiver sets the backup link as the second priority, while the main link remains operational to ensure smooth connection.

[0121] T begins (14:00:07). The mining truck enters the obstruction zone of RS2, the main link signal is blocked, the receiver seamlessly switches to the backup link, and the signal strength stabilizes at -73dBm without interruption. Subsequent operations: According to the scheduling logic of RS1, the T end and T recovery operations are executed in sequence, RS2 returns to its initial state and goes into sleep mode.

[0122] Phase 3: Multi-metasurface relay scheduling.

[0123] The mine car continues its journey, and the system triggers the scheduling execution of metasurfaces such as RS3 and RS4 sequentially according to the scheduling sequence table. The scheduling process of each metasurface is consistent with that of RS1 and RS2, forming a relay-like process where the mine car moves and the metasurface follows. There is a 0.5-second overlap between the scheduling of adjacent metasurfaces. For example, T recovery of RS1 is 14:00:07.5, and T start of RS2 is 14:00:07. During the overlap period, the system only keeps the currently active alternative link working and shuts down other links to avoid multipath interference.

[0124] Furthermore, when the status of the mining truck changes, the scheduling timetable is adjusted to adapt to the status of the mining truck.

[0125] Specifically, changes in the mine car's status include acceleration / deceleration and temporary stops. If the mine car accelerates, i.e., its speed increases from 10 km / h to 15 km / h, real-time channel feedback data shows that the signal attenuation rate is accelerating. The system recalculates the scheduling timing table and advances the T warning by 0.3 seconds to ensure sufficient adjustment time. If the mine car decelerates, the T warning is delayed to avoid premature adjustment and resource waste. If the mine car temporarily stops in the metasurface obstruction area, the production management system reports "operating status = stopped." The system extends the activation time of the alternative link to maintain signal stability until the mine car restarts.

[0126] In addition to changes in the status of the mine car, there is also the possibility of route changes. If the mine car scheduling route is temporarily changed, such as being diverted to an auxiliary roadway, the system immediately stops the original scheduling timetable, calls the model again to predict the occlusion time point of the supersurface on the new route, and generates a new scheduling timetable to avoid misscheduling.

[0127] Meanwhile, there is also the possibility of channel state changes, including interference and signal attenuation. Regarding interference, if the real-time channel feedback sensor detects electromagnetic interference, such as a pantograph arc causing a sudden SNR drop of <10dB, the system immediately adjusts the angle of the adjustable metasurface, such as increasing the pitch angle by 5° to avoid the direction of interference. Simultaneously, it switches to a frequency band with stronger anti-interference capabilities, such as switching from 2.4G to 5.8G, ensuring a signal strength ≥-80dBm. Regarding signal attenuation, if changes in the tunnel environment (such as increased dust) exacerbate signal attenuation, the system increases the number of metasurfaces in the alternative links, such as expanding from 1 adjustable metasurface + 1 fixed metasurface to 1 adjustable metasurface + 2 fixed metasurfaces, compensating for signal attenuation through multiple reflection paths.

[0128] During operation, the system automatically records execution data for each scheduling operation, including mine car data, key occlusion time points, angle adjustment parameters, and signal quality indicators (RSRP, SNR, and interruption status), forming an iterative dataset. Every 10,000 iterations, the ground server initiates iterative model training, for example, fine-tuning the learning rate to 0.0001 and the number of training rounds to 20, optimizing the model's prediction accuracy, and adapting to new operating scenarios, such as changes in mine car speed and tunnel environment. After training is complete, the optimized model parameters and scheduling strategy are distributed to the underground edge gateway, replacing the old version and continuously improving system performance.

[0129] Experimental instructions

[0130] I. Experimental Scenario Design.

[0131] 1. Basic scene parameters.

[0132] Tunnel type: Long straight underground coal mine tunnel, 1000m long, 5m wide, and 3.5m high. The tunnel walls are made of concrete, and the electromagnetic attenuation coefficient is 0.25dB / m (2.4G frequency band).

[0133] Communication requirements: The base station is deployed at 0m in the alley, and the receiver is deployed at 1000m. The communication thresholds must be RSRP≥-80dBm and SNR≥15dB.

[0134] Mining car scenarios: covering 3 typical scenarios - ① Conventional coal car (15m long, 2.5m wide); ② Large equipment car (25m long, 3m wide); ③ Parallel driving of two cars (two conventional coal cars driving side by side, with a total width of 5m).

[0135] Power supply conditions: There is one convenient power supply point every 300m in the tunnel, that is, there are a total of 4 convenient power supply points, located at 0m, 300m, 600m and 900m respectively.

[0136] 2. Control of experimental variables.

[0137] Communication frequency band: The 2.4G frequency band is used uniformly.

[0138] Metasurface core parameters: The passive metasurface unit parameters of this application and the fixed bypass scheme are the same. The angle adjustment range of the adjustable metasurface is ±30° with an accuracy of 0.1°.

[0139] Test tools: mine signal tester (sampling frequency 10Hz, RSRP measurement accuracy ±2dBm), laser rangefinder, time sequence data recorder.

[0140] II. Experimental Procedure.

[0141] (I) Deployment phase of the plan.

[0142] 1. Deployment of a fixed bypass link solution.

[0143] Link design: Based on the maximum coverage size of the mining truck + the worst-case scenario of two trucks running in parallel, the fixed bypass link is planned from the base station → top metasurface → side metasurface → receiver. The transmission path length is 1500m, which is 50% longer than the straight path.

[0144] Metasurface deployment: Eight fixed metasurfaces are deployed along a fixed link at locations of 100m, 250m, 400m, 550m, 700m, 800m, 850m, and 950m.

[0145] 2. Deployment of this method.

[0146] Link design: In a normal scenario, the main link is base station → passive metasurface → receiver, with a straight path length of 1000m. In extreme scenarios, the alternative link is dynamically switched based on the timing of the mine truck's obstruction.

[0147] Metasurface deployment: 3 sets of adjustable metasurfaces (deployed at convenient power supply points at 300m, 600m, and 900m) + 6 sets of fixed metasurfaces (deployed at 150m, 450m, 500m, 750m, 850m, and 950m). The fixed metasurfaces can be adapted to multiple primary and backup links.

[0148] (ii) Signal quality test.

[0149] 1. Standard scenario test: In this scenario, there are no mining trucks or only small mining trucks passing through.

[0150] Test method: Continuously collect RSRP and SNR data from the receiving end, and record the average value, minimum value, and fluctuation range within 1 hour;

[0151] Scenario breakdown: ① No mining trucks pass through; ② Small maintenance vehicles (8m long, 2m wide) travel along the track without obstructing the main link.

[0152] 2. Extreme scenario test, in which mining trucks obstruct the view.

[0153] Test method: Three types of mining trucks were driven at a speed of 10km / h. The number of signal interruptions, the duration of each interruption, and the signal stability value were recorded when the mining trucks passed through the 0-1000m section.

[0154] Data acquisition: Record one set of signal data every 50m, and synchronously record the metasurface scheduling actions (this method).

[0155] (III) Statistics on the number of metasurfaces.

[0156] Initial deployment quantity: The number of fixed / adjustable metasurfaces for both schemes is counted separately.

[0157] Quantity after scene change: Simulate two types of scene changes: ① The length of the mine car increases from 15m to 25m; ② The tunnel excavation extends from 200m to 1200m. Calculate the number of new metasurfaces required for the scheme.

[0158] III. Comparison of Experimental Results

[0159] 1. Signal quality comparisons are shown in Table 2.

[0160] Table 2 Signal Quality Comparison

[0161]

[0162] 2. The number of metasurfaces used is shown in Table 3.

[0163] Table 3 Comparison of Metasurface Usage Quantities

[0164]

[0165] IV. Experimental Conclusions.

[0166] Signal quality: In normal scenarios, the signal strength of this method (average -69.5~-70.2dBm) is significantly better than the fixed bypass link scheme (-77.3~-78.1dBm), far from the communication threshold, and there is no weak signal lag. In extreme scenarios, this method is uninterrupted, while the fixed scheme has insufficient link design redundancy, resulting in 3 interruptions when two vehicles are running in parallel, with a maximum interruption duration of 500ms, and the signal fluctuation amplitude is 2 to 3 times that of this method.

[0167] Number of metasurfaces: The initial deployment of this method is slightly more (+11.1%), but the fixed solution cannot reuse metasurfaces and needs to continuously add more after the scenario changes (44.4% more in 3 years). In contrast, this method uses fixed metasurfaces for multi-link reuse, so the number remains stable.

[0168] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0169] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for installing a passive metasurface in a coal mine underground roadway, characterized in that, include: Collect static environmental data of underground coal mine roadways, timing data of mine car operation, and timing data of signal transmission in roadways at different time periods; A three-dimensional mesh model of the tunnel is established based on the static environmental data of the tunnel. The timing data of the mine car operation and the timing data of the signal transmission are imported into the three-dimensional mesh model of the tunnel to form a timing-series three-dimensional model. Based on the timing patterns of mine car occlusion and signal transmission, a conventional main link and a timing-sequential alternative link are planned in the timing-sequential 3D model. The conventional main link is used during periods without occlusion, and the timing-sequential alternative link is used during periods of mine car occlusion. Multiple passive metasurfaces are deployed in the tunnel according to the conventional main link and the time-sequential alternative link. The passive metasurfaces include adjustable metasurfaces and fixed metasurfaces. The adjustable metasurfaces can adjust the orientation angle, and the fixed metasurfaces have a fixed orientation angle. Real-time data of mining trucks is obtained from the mine production management system. The real-time data of mining trucks is input into the time-series scheduling model to predict the key time points of shading of each of the adjustable metasurfaces and generate a scheduling time series table according to the direction of travel of the mining trucks. According to the scheduling timetable, each of the adjustable metasurfaces is controlled to adjust its orientation angle in sequence to switch between the conventional main link and the time-sequential alternative link.

2. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, The static environmental data of the tunnel includes three-dimensional point cloud data of the tunnel, key cross-sectional dimensions, tunnel wall material, obstacle distribution, location and curvature of bends / intersections, and coordinates of power supply convenience points. After collecting the static environmental data of the tunnel, data processing is performed, which includes point cloud denoising, registration and stitching, segmentation simplification, and static data standardization.

3. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, After collecting the static data of the tunnel environment, the running time data of the mine car, and the signal transmission time data, coordinate unification and time synchronization processing are performed to form a time-series associated dataset. Based on the time-series associated dataset, the three-dimensional mesh model of the tunnel and the time-series three-dimensional model are established.

4. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, The planning principles for the conventional main link are to minimize signal attenuation and the number of reflections. The planning principles for the time-sequential alternative links include timing adaptation, redundancy coverage, and signal compliance. The timing adaptation principle ensures that each alternative link in the time-sequential alternative links corresponds to a set of mine car-metasurface occlusion timing pairs, so that when a mine car occludes a metasurface in the link, the corresponding alternative link can be activated before the occlusion arrives. The redundancy coverage principle ensures that at least one alternative link is planned for each critical metasurface of the conventional main link, and that the alternative link has no risk of overlapping occlusion with the conventional main link. The signal compliance principle ensures that the transmission distance of the alternative link is less than or equal to 30% longer than that of the conventional main link, and the signal attenuation increase is less than or equal to 5dB.

5. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, Before deploying the passive metasurface, deployment locations are first selected in the tunnel. The selection criteria include timing compatibility, power supply feasibility, and link compatibility. Then, the initial angle of the passive metasurface is calculated. Finally, the passive metasurface is deployed according to the deployment location and the initial angle.

6. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, The time-series scheduling model is built on an LSTM+attention mechanism and has been trained on a training dataset.

7. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, The time-series scheduling model is deployed in the downhole edge gateway and TensorRT is used for model acceleration.

8. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, After predicting the critical time point of the occlusion, the critical time point of the occlusion is corrected by combining the feedback data of the real-time channel.

9. The method for installing a passive metasurface in an underground coal mine roadway according to claim 1, characterized in that, When the status of a mining truck changes, the scheduling timetable is adjusted to adapt to the status of the mining truck.

Images