Coal mine underground trackless rubber-tyred vehicle-roadway collaborative sensing system and method

Abstract

The invention discloses a coal mine underground trackless rubber-tyred vehicle-lane collaborative sensing system and method, and relates to the technical field of coal mine underground transportation, and the system comprises a vehicle end sensing module which comprises a vehicle end multi-line laser radar, an inertia measurement unit and a vehicle end millimeter wave radar which are disposed on a trackless rubber-tyred vehicle; the roadway side sensing module comprises roadway side fixed multi-line laser radars arranged at key nodes of an underground roadway; the time synchronization module comprises UWB units arranged at the vehicle end and the lane side respectively, and the UWB units are configured to interact synchronization pulses between the vehicle end and the lane side, so that data collected by the vehicle end sensing module and the lane side sensing module have synchronous timestamps; and the data processing module is configured to fuse data acquired by the vehicle end sensing module and the lane side sensing module based on the synchronous timestamps to generate global positioning information of the trackless rubber-tyred vehicle. The problem that the performance of the sensor is reduced due to factors such as positioning drifting, blind area sensing deficiency and dust of a single underground sensor is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of underground coal mine transportation technology, and in particular to a vehicle-roadway collaborative sensing system and method for trackless rubber-tired vehicles in underground coal mines. Background Technology

[0002] Underground auxiliary transportation in coal mines is a crucial link in coal mine production. Trackless rubber-tired vehicles, as explosion-proof transportation equipment that can travel without tracks, are widely used for the efficient transfer of personnel, materials, and large equipment underground, and have become the core equipment of the coal mine auxiliary transportation system.

[0003] Traditional trackless rubber-tired vehicles rely on dedicated drivers. Due to the narrowness of underground roadways, insufficient lighting, complex environment, and monotonous operating routes, drivers are prone to inattention or fatigue during long working hours, increasing the risk of traffic accidents. The introduction of unmanned driving technology provides an effective way to improve transportation safety and efficiency. Currently, the perception solutions for unmanned trackless rubber-tired vehicles in coal mines mainly have the following shortcomings: First, in the harsh environment of low light and high dust underground, a single vision sensor is prone to positioning drift due to the failure of feature point extraction, resulting in poor environmental adaptability and difficulty in meeting the needs of continuous and stable perception.

[0004] Secondly, while existing technologies using ordinary lidar can overcome light interference, they can only acquire limited ambient point cloud information around the vehicle. They cannot cover blind spots on the side of the lane, such as behind curves at intersections and at the top of slopes. They also lack collaborative perception with fixed facilities in the lane (such as signs and signal devices), which can easily lead to a disconnect between vehicle and lane information and cause perception blind spots and misjudgments in complex lane conditions.

[0005] To address these issues, existing solutions attempt to integrate multiple vehicle-mounted sensors, but they lack systematic optimization for specific working conditions such as high dust levels, high humidity, and the absence of satellite positioning in underground environments. For example, the installation layout of the lidar fails to effectively mitigate the impact of dust adhesion on laser penetration; the lack of a high-precision time synchronization mechanism between the vehicle-mounted and roadway-side sensing devices leads to inconsistencies in the timing of multi-source data. Furthermore, existing solutions lack effective sensing redundancy and degradation strategies in the event of sensor failure, making it difficult to maintain sufficient sensing accuracy and robustness when some sensors fail.

[0006] It is evident that existing sensing technologies are insufficient to fully meet the collaborative sensing requirements of unmanned trackless rubber-tired vehicles in coal mines for high precision, full coverage, and strong robustness, especially in areas such as blind zone coverage on the roadway side, spatiotemporal synchronization of multi-source data, dust interference suppression, and sensor fault tolerance.

[0007] Therefore, there is an urgent need for an unmanned underground sensing system and method that can enable collaborative work between vehicle-side and roadway-side sensing devices, possess high-precision spatiotemporal synchronization capabilities, and support redundant and differentiated data transmission of sensing data, so as to ensure the safe and stable operation of unmanned trackless rubber-tired vehicles in complex and harsh underground environments. Summary of the Invention

[0008] In a first aspect, to solve the above-mentioned technical problems, this invention provides a trackless rubber-tired vehicle-roadway cooperative sensing system for underground coal mines, comprising: The vehicle-side perception module includes a multi-line lidar, an inertial measurement unit, and a millimeter-wave radar installed on the trackless rubber-tired vehicle. The roadway-side sensing module includes at least one roadway-side fixed multi-line lidar installed at key nodes in the underground roadway; A time synchronization module includes UWB units respectively located at the vehicle end and the alleyside. The UWB units are configured to exchange synchronization pulses between the vehicle end and the alleyside, so that the data collected by the vehicle end sensing module and the alleyside sensing module have synchronized timestamps. The data processing module is configured to generate global positioning information of the trackless rubber-tired vehicle by fusing data collected by the vehicle-side sensing module and the alley-side sensing module based on the synchronized timestamp.

[0009] Furthermore, the alleyway-side perception module also includes an infrared camera coaxially mounted with the alleyway-side fixed multi-line lidar, the infrared camera being used to collect image semantic information of the key nodes; the data processing module is also configured to fuse the image semantic information.

[0010] Furthermore, the data processing module includes an on-board domain controller and a ground dispatch server; The vehicle-mounted domain controller is configured to fuse the mileage data from the vehicle-mounted multi-line lidar with the attitude data from the inertial measurement unit to obtain the preliminary positioning result of the trackless rubber-wheeled vehicle. The ground dispatch server is configured to use a factor graph optimization algorithm to fuse and optimize the preliminary positioning results with the data collected by the alleyway side sensing module to obtain the global positioning information.

[0011] Furthermore, the ground dispatch server is also configured to introduce the absolute coordinates of the benchmark mileage markers set in the roadway as constraints during the fusion optimization process of the factor graph optimization algorithm.

[0012] Furthermore, the data processing module is also configured to: Monitor the point cloud data quality of the vehicle-mounted multi-line lidar; In response to the point cloud data quality being lower than a first threshold, the weight of the vehicle-mounted multi-line lidar data in the fusion is reduced, and the weight of the alley-side fixed multi-line lidar data in the fusion is increased.

[0013] Furthermore, the data processing module is also configured to: In response to the point cloud data quality being lower than the first threshold, the vehicle-mounted millimeter-wave radar is controlled to expand its lateral detection angle to enhance lateral perception.

[0014] Furthermore, it also includes a communication module, which comprises mining 5G base stations deployed at intervals along the underground roadways; The system is configured to transmit point cloud data acquired by a multi-line lidar via UDP protocol, and transmit data acquired by the inertial measurement unit and the vehicle-mounted millimeter-wave radar via TCP protocol in a 5G control channel.

[0015] Furthermore, the system is configured as follows: In response to the alley-side sensing module detecting a sudden obstacle, the 5G network slicing technology is triggered to establish a high-priority transmission channel for the sensing data related to the sudden obstacle.

[0016] In a second aspect, the present invention provides a method for vehicle-roadway cooperative sensing of a trackless rubber-tired vehicle in a coal mine based on the aforementioned system, comprising the following steps: The vehicle-side UWB unit and the alley-side UWB unit interact with each other to synchronize pulses, and the vehicle-side sensing data and the alley-side sensing data are time-stamped and synchronized. By integrating the vehicle-side sensing data synchronized with the alley-side sensing data after time stamping, global positioning information of the trackless rubber-wheeled vehicle is generated.

[0017] Furthermore, the fusion of time-stamped vehicle-side sensing data and lane-side sensing data specifically includes: In the vehicle domain controller, the mileage data from the vehicle-mounted multi-line lidar and the attitude data from the inertial measurement unit are fused by Kalman filtering to obtain preliminary positioning results; In the ground dispatch server, the preliminary positioning results and the alleyside sensing data are fused and optimized using a factor graph optimization algorithm to obtain the global positioning information.

[0018] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: This invention fundamentally solves the time alignment problem of multi-source heterogeneous data between the vehicle and the roadway by introducing a cross-domain time synchronization mechanism based on UWB pulse interaction, establishing a unified time series benchmark for high-precision data fusion. Based on this benchmark, a two-level positioning architecture of "vehicle-side KF local fusion + ground-side FGO global optimization" and a redundant perception strategy of "roadway-side blind spot filling + dynamic weight adjustment" are combined to effectively overcome the problems of single-sensor positioning drift, blind spot perception loss, and sensor performance degradation caused by dust and other factors underground. This improves global positioning accuracy to the centimeter level and enhances the overall robustness of the system. Simultaneously, by designing differentiated UDP / TCP transmission strategies for perception data with different characteristics such as point clouds and IMUs, and utilizing 5G slicing technology to ensure low-latency transmission of bursty high-priority data, a balance between "low latency" and "high reliability" is achieved in the complex underground network environment. This provides stable, accurate, and fully covered real-time environmental perception information for unmanned driving decision-making and control, comprehensively ensuring the driving safety and transportation efficiency of trackless rubber-tired vehicles in coal mines under harsh working conditions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the overall system structure disclosed in this invention; Figure 2 This is a schematic diagram of the installation layout of each sensor in the vehicle-side perception module disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of the installation layout of the core components at the turning node of the alleyway-side sensing module disclosed in an embodiment of the present invention. Figure 4 This is a flowchart of the system startup phase disclosed in this invention; Figure 5 This is a flowchart of the normal driving phase of the system disclosed in this invention; Figure 6 This is a flowchart illustrating the processing of special scenarios in the system disclosed in this invention, specifically the flowchart for a vehicle meeting scenario; Figure 7 This is a flowchart illustrating the processing of a special scenario in the system disclosed in this invention, showing the flowchart when any sensor fails.

[0021] In the picture: 11. 16-line intrinsically safe lidar for mining; 12. Instrument panel; 13. Vehicle-mounted communication terminal; 14. Millimeter-wave radar for mining; 15. 8-line intrinsically safe lidar for mining; 16. Vehicle-mounted domain controller; 21. Roadway side control unit; 22. Mine infrared thermal imaging camera; 23. 8-line intrinsically safe fixed lidar for mining; 24. Mine 5G base station; 25. Roadway wall; 26. Roadway; 27. Roadway side communication terminal. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 This invention provides a vehicle-roadway collaborative sensing system for trackless rubber-tired vehicles in coal mines. The system aims to achieve full coverage and highly robust sensing of the operating environment of trackless rubber-tired vehicles underground through the collaboration of vehicle-side and roadway-side sensing devices, high-precision spatiotemporal synchronization, and intelligent data fusion. It mainly includes a vehicle-side sensing module, a roadway-side sensing module, a time synchronization module, a communication module, and a data processing module. The following provides a detailed description of each module.

[0024] First, a detailed explanation of the vehicle-side perception module will be provided.

[0025] Please see Figure 2 The vehicle-mounted sensing module is installed on the unmanned trackless rubber-tired vehicle to collect real-time information on the vehicle's own status and the surrounding near-field environment. Its core components include a multi-line lidar, an inertial measurement unit, and a millimeter-wave radar installed on the trackless rubber-tired vehicle.

[0026] The vehicle-mounted multi-line lidar system consists of a 16-line intrinsically safe lidar 11 (primary radar) and two 8-line intrinsically safe lidars 15 (auxiliary radars). The 16-line intrinsically safe lidar 11 is fixed to the center of the front of the vehicle roof using explosion-proof brackets, with a horizontal field of view covering at least 120° in front and its optical axis parallel to the vehicle's longitudinal axis. The two auxiliary 8-line intrinsically safe lidars 15 are fixed to both sides of the vehicle body using side-mounted explosion-proof brackets, installed at a height of 1.5-2.0m above the tunnel floor, with each unit having a horizontal field of view covering at least 90°. The three systems work together to achieve 360° three-dimensional point cloud acquisition around the vehicle without blind spots.

[0027] In this design, a mining millimeter-wave radar 14 is embedded inside the front bumper. Its installation height is 0.8-1.0m above the tunnel floor. Its antenna orientation is basically the same as that of the 16-line intrinsically safe mining lidar 11. It is mainly used to detect targets that are not sensitive to laser reflection characteristics, such as transparent glass covers and highly reflective metal brackets, as an effective supplement to lidar.

[0028] The inertial measurement unit (IMU) is not an independent external device, but is highly integrated inside the vehicle-mounted domain controller 16. The vehicle-mounted domain controller 16 itself is fixed to a dedicated explosion-proof box at the rear of the cab with explosion-proof bolts, and its installation position is designed to coincide with the vehicle's center of gravity axis in order to minimize the interference of vehicle body bumps on the accuracy of IMU measurements.

[0029] Next, the alley-side sensing module will be described in detail.

[0030] Please see Figure 3 The roadway-side sensing module is fixedly deployed at key topological nodes in the underground roadway to compensate for the inherent blind spots of vehicle-mounted sensors and provide global, prior environmental information. It is deployed at key locations such as roadway intersections, the start and end points of slopes, and the middle sections of long straight roadways exceeding 1 km in length. The roadway-side sensing module includes at least one fixed multi-line lidar unit located at key nodes in the underground roadway. Specifically, each key node in the roadway typically includes an 8-line intrinsically safe fixed lidar unit 23, a mining infrared thermal imaging camera 22, and a roadway-side control unit 21 (LCU).

[0031] The 8-line intrinsically safe fixed lidar 23 for mining is installed on the sidewall of the roadway using an L-shaped explosion-proof bracket, with an installation height of 2.5-3.0m from the floor. In particular, at the junction of roadways, two 8-line intrinsically safe fixed lidar 23s for mining need to be deployed and installed opposite each other at an angle of 50°-70° to achieve complete coverage of the intersection blind zone.

[0032] Among them, the mining infrared thermal imaging camera 22 and the 8-line intrinsically safe fixed lidar 23 of the same node are coaxially installed, with the distance between them controlled within 0.2-0.5m, and the field of view of the lens is matched to ensure that the acquired infrared image and the lidar point cloud can be accurately aligned in space, providing rich semantic information for the perception data.

[0033] The roadway-side control unit 21 is installed in an explosion-proof box on the roadway wall 25, 1.0-1.5m below the 8-line intrinsically safe fixed lidar 23 at the same node. It is connected to the 8-line intrinsically safe fixed lidar 23 and the mining infrared thermal imaging camera 22 via a mine-use intrinsically safe shielded cable with a length not exceeding 8m. It is responsible for local preprocessing of the raw sensing data, such as denoising and coordinate transformation of point clouds, and grayscale enhancement of infrared images to adapt to the low-light environment of 1 lux to 5 lux underground.

[0034] Secondly, the time synchronization module will be explained in detail.

[0035] The time synchronization module includes UWB units respectively set at the vehicle end and the roadside. The UWB units are configured to exchange synchronization pulses between the vehicle end and the roadside so that the data collected by the vehicle end sensing module and the roadside sensing module have synchronized timestamps.

[0036] The vehicle-side UWB unit is integrated inside the vehicle-mounted communication terminal 13. The alley-side UWB unit is integrated inside the explosion-proof box of the alley-side control unit 21, with one deployed at each key node as a "time anchor point" for that area. The vehicle-side and alley-side UWB units exchange synchronization pulse signals wirelessly to achieve microsecond-level time synchronization, thereby giving all data frames collected by the vehicle-side and alley-side units a unified and accurate timestamp.

[0037] Furthermore, the communication module is described in detail.

[0038] The communication module includes 5G base stations 24 deployed at intervals along the underground roadway. These 5G base stations 24 are deployed at 300-500m intervals along the main underground transport roadway 26, fixed to the sidewall 3.0-3.5m from the floor using explosion-proof brackets, and employ directional antennas to optimize signal coverage. The vehicle-mounted communication terminal 13 and the roadway-side communication terminal 27 serve as access points for the vehicle and roadway sides, respectively, connecting to the 5G network and responsible for reliable, low-latency transmission of sensing data and control commands.

[0039] The system transmits point cloud data collected by the multi-line lidar via the UDP protocol, and transmits data collected by the inertial measurement unit and the vehicle-mounted millimeter-wave radar via the TCP protocol in the 5G control channel.

[0040] Finally, the data processing module will be described in detail.

[0041] The data processing module is configured to generate global positioning information for the trackless rubber-tired vehicle by integrating data collected from the vehicle-side perception module and the alley-side perception module based on synchronized timestamps. Specifically, the data processing module adopts a two-level processing architecture that is "edge-cloud" collaborative, which includes an on-board domain controller 16 and a ground dispatch server to balance real-time performance and global optimization.

[0042] The vehicle-mounted domain controller 16 is responsible for real-time acquisition, filtering, and initial fusion of mileage data from the vehicle-mounted multi-line lidar and attitude data from the inertial measurement unit to obtain preliminary positioning results for the trackless rubber-tired vehicle. The ground dispatch server has stronger computing power and, based on the factor graph optimization algorithm, is responsible for fusing global data from multiple vehicles and multiple laneside nodes, performing optimization calculations and decision planning to obtain global positioning information.

[0043] The following combination Figures 4 to 7 The flowchart details the system's workflow.

[0044] Please see Figure 4 The flowchart shown illustrates the system startup and initialization phases. After the trackless rubber-wheeled vehicle is powered on, each vehicle-end sensor initiates a self-test. Once the status is normal, it sends a "ready" signal to the onboard domain controller 16. Simultaneously, the roadway-side control unit 21 establishes a connection with the ground dispatch server via the 5G network and reports the equipment status. Subsequently, the vehicle travels to a straight roadway at least 500m long for initial calibration: the vehicle-end lidar collects 25 point clouds from the roadway wall and matches them with the pre-stored 3D model to complete initial pose calibration; the IMU performs static zero-bias calibration. The roadway-side control unit 21 controls the 8-line intrinsically safe fixed lidar 23 to scan nearby benchmark mileage markers, comparing the identified coordinates with the pre-stored absolute coordinates to complete the roadway-side coordinate system calibration. At this time, the vehicle-end UWB unit establishes a connection with the nearest roadway-side UWB unit to complete high-precision time synchronization. The ground dispatch server sends the initial driving trajectory containing global coordinates to the vehicle, and the system enters the ready state.

[0045] Please see Figure 5 The flowchart illustrates the normal driving and collaborative perception phases. During driving, the vehicle-side perception module continuously operates—point cloud data collected by the main and auxiliary LiDARs are stitched together into a full-circle point cloud map in the vehicle domain controller 16, and dust noise is filtered out; the IMU outputs the vehicle attitude in real time; and the vehicle-side millimeter-wave radar outputs information about targets ahead. The vehicle domain controller 16 uses a Kalman filter (KF) algorithm to fuse LiDAR odometer data and IMU data to generate preliminary vehicle positioning results and a local environment model, which are then uploaded via the 5G network.

[0046] Meanwhile, the roadway-side sensing module monitors the blind areas it covers—an 8-line intrinsically safe fixed lidar 23 collects point clouds of the blind areas, and a mine-use infrared thermal imaging camera 22 collects images, which are then semantically segmented and identified by the roadway-side control unit 21 to determine the target type (such as pedestrians or vehicles). After preprocessing the data and performing global coordinate transformation, the roadway-side control unit 21 attaches a UWB synchronization timestamp and uploads it to the ground dispatch server via the 5G network.

[0047] After receiving the initial positioning data from the vehicle and the perception data from the roadside, the ground dispatch server performs global fusion—using the Factor Graph Optimization (FGO) algorithm to optimize the vehicle data, roadside data, and the absolute coordinates of the benchmark kilometer markers in the road (e.g., (1200,0,0)) as constraints. This outputs global positioning information with centimeter-level accuracy and constructs a vehicle-road integrated global environment model. Subsequently, based on this model and driving plan, the dispatch center issues specific control commands to the vehicle (e.g., "speed 15km / h, keep to the right"), which the vehicle receives and executes via the CAN bus.

[0048] Specifically, to balance transmission latency and reliability, the communication module employs a differentiated strategy. Non-critical, large-volume point cloud data is transmitted using the UDP protocol to ensure low latency. Critical IMU attitude data and millimeter-wave radar target data, however, are transmitted via the more reliable TCP protocol, embedded in the 5G control channel with priority. All data frames are bound to a unified timestamp generated by the UWB unit.

[0049] Please see Figure 6 The flowchart illustrates a specific scenario of vehicles meeting at a fork in the roadway. As the vehicle approaches the fork, the intrinsically safe fixed lidar 23 deployed on the roadway side detects the approaching vehicle. The roadway side control unit 21 immediately prioritizes the vehicle's position, speed, and other information, using 5G network slicing technology to seize bandwidth and upload it to the dispatch center with extremely low latency. After verifying the data from multiple sources, the dispatch center simultaneously issues commands to both the vehicle and the oncoming vehicle to "reduce to 8 km / h" and "keep to the right," until both vehicles safely pass the fork and resume normal speed.

[0050] Please see Figure 7 The flowchart illustrates a specific scenario involving a sensor failure. This demonstrates the perceptual fault tolerance and redundancy of the present invention. When the system detects that a point cloud loss rate exceeding a threshold (e.g., 30%) due to dust obstruction by a vehicle-mounted sensor (such as the right-side auxiliary 8-line intrinsically safe mining lidar 15), the vehicle-mounted domain controller 16 sends a fault alarm. Upon receiving the alarm, the ground dispatch server dynamically adjusts the data fusion weights: reducing the weight of the faulty radar from 5 to 1, while simultaneously increasing the weight of the roadway-side 8-line intrinsically safe mining lidar 23 covering the blind spot from 5 to 10. Furthermore, the system activates the lateral enhancement detection mode of the vehicle-mounted mining millimeter-wave radar 14, expanding its right-side detection angle. Through this redundancy strategy of reducing fault weights, increasing roadway-side weights, and activating backup radars, the system can maintain sufficient positioning accuracy and obstacle recognition rate during single-point failures, ensuring vehicle safety.

[0051] Furthermore, based on the above system, this invention also provides a method for vehicle-roadway cooperative sensing of trackless rubber-tired vehicles in coal mines, which mainly includes the following steps: S1. The vehicle-side UWB unit and the roadside UWB unit interact with each other via synchronization pulses to timestamp the vehicle-side sensing data and the roadside sensing data.

[0052] S2. Integrate the vehicle-side perception data synchronized with the alley-side perception data after timestamp to generate global positioning information for the trackless rubber-wheeled vehicle.

[0053] Specifically, the fusion of vehicle-side perception data and lane-side perception data after time-stamp synchronization includes: First, in the vehicle-mounted domain controller, the mileage data of the vehicle-mounted multi-line LiDAR and the attitude data of the inertial measurement unit are fused through Kalman filtering to obtain preliminary positioning results; then, in the ground dispatch server, the preliminary positioning results and lane-side perception data are fused and optimized through a factor graph optimization algorithm to obtain global positioning information.

[0054] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines, characterized in that, include: The vehicle-side perception module includes a multi-line lidar, an inertial measurement unit, and a millimeter-wave radar installed on the trackless rubber-tired vehicle. The roadway-side sensing module includes at least one roadway-side fixed multi-line lidar installed at key nodes in the underground roadway; The time synchronization module includes UWB units respectively set at the vehicle end and the alley side. The UWB units are configured to exchange synchronization pulses between the vehicle end and the alley side so that the data collected by the vehicle end sensing module and the alley side sensing module have synchronized timestamps. as well as The data processing module is configured to generate global positioning information of the trackless rubber-tired vehicle by fusing data collected by the vehicle-side sensing module and the alley-side sensing module based on the synchronized timestamp.

2. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 1, characterized in that, The alleyway-side perception module also includes an infrared camera coaxially mounted with the alleyway-side fixed multi-line lidar. The infrared camera is used to collect image semantic information of the key nodes. The data processing module is also configured to fuse the image semantic information.

3. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 1, characterized in that, The data processing module includes an on-board domain controller and a ground dispatch server; The vehicle-mounted domain controller is configured to fuse the mileage data from the vehicle-mounted multi-line lidar with the attitude data from the inertial measurement unit to obtain the preliminary positioning result of the trackless rubber-wheeled vehicle. The ground dispatch server is configured to use a factor graph optimization algorithm to fuse and optimize the preliminary positioning results with the data collected by the alleyway side sensing module to obtain the global positioning information.

4. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 3, characterized in that, The ground dispatch server is also configured to introduce the absolute coordinates of the benchmark kilometer markers set in the tunnel as constraints during the fusion optimization process of the factor graph optimization algorithm.

5. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 1, characterized in that, The data processing module is also configured to: Monitor the point cloud data quality of the vehicle-mounted multi-line lidar; In response to the point cloud data quality being lower than a first threshold, the weight of the vehicle-mounted multi-line lidar data in the fusion is reduced, and the weight of the alley-side fixed multi-line lidar data in the fusion is increased.

6. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 5, characterized in that, The data processing module is also configured to: In response to the point cloud data quality being lower than the first threshold, the vehicle-mounted millimeter-wave radar is controlled to expand its lateral detection angle to enhance lateral perception.

7. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 1, characterized in that, It also includes a communication module, which comprises mining 5G base stations deployed at intervals along the underground roadways; The system is configured to transmit point cloud data acquired by a multi-line lidar via UDP protocol, and transmit data acquired by the inertial measurement unit and the vehicle-mounted millimeter-wave radar via TCP protocol in a 5G control channel.

8. The trackless rubber-tired vehicle-roadway collaborative sensing system for underground coal mines according to claim 7, characterized in that, The system is configured as follows: In response to the alley-side sensing module detecting a sudden obstacle, the 5G network slicing technology is triggered to establish a high-priority transmission channel for the sensing data related to the sudden obstacle.

9. A method for collaborative sensing between a trackless rubber-tired vehicle and a roadway in a coal mine, based on the system described in any one of claims 1-8, characterized in that, Includes the following steps: The vehicle-side UWB unit and the alley-side UWB unit interact with each other to synchronize pulses, and the vehicle-side sensing data and the alley-side sensing data are time-stamped and synchronized. By integrating the vehicle-side sensing data synchronized with the alley-side sensing data after time stamping, global positioning information of the trackless rubber-wheeled vehicle is generated.

10. The method for coordinated perception between trackless rubber-tired vehicles and roadways in underground coal mines according to claim 9, characterized in that, The fusion of vehicle-side sensing data and lane-side sensing data after timestamp synchronization specifically includes: In the vehicle domain controller, the mileage data from the vehicle-mounted multi-line lidar and the attitude data from the inertial measurement unit are fused by Kalman filtering to obtain preliminary positioning results; In the ground dispatch server, the preliminary positioning results and the alleyside sensing data are fused and optimized using a factor graph optimization algorithm to obtain the global positioning information.

Images