An unmanned transport vehicle system and operation method applicable to phosphate mining roadways

Abstract

The present invention provides an unmanned transport vehicle system and an operation method applicable to a phosphate ore mining roadway, including an execution system for carrying the entire transport vehicle system, the execution system including an unmanned transport vehicle; a sensing system that relies on sensors to provide environmental information for the unmanned transport vehicle of the execution system to assist the unmanned transport vehicle in completing navigation and positioning; a control system that includes two modules, a decision-making module and a planning module, which uses the environmental information from the sensing system to obtain control signals for the unmanned transport vehicle; the execution system receives the control signals issued by the control system and sends execution information such as direction, speed, acceleration, heading angle, and driving trajectory to the unmanned transport vehicle, enabling the unmanned transport vehicle to avoid local obstacles and track the global path, thereby completing the unmanned driving task in the phosphate ore mining roadway.

Description

Technical Field

[0001] The present invention belongs to the technical field of phosphate ore mining equipment, and particularly relates to an unmanned transport vehicle system applicable to phosphate ore mining roadways. Background Art

[0002] Phosphate ore mining roadways are passages for the mining and transportation of phosphate ore, allowing mining transport vehicles to pass through. The light in the roadways is dim, and the working environment is dusty. Working in such an environment for a long time is likely to affect the physical health of the staff. Therefore, the use of unmanned mining and transportation equipment can reduce the number of underground workers, which is beneficial to improving work efficiency and the safety of mining and transportation. Currently, with the development of robot technology, mobile robots based on scene modeling, scene recognition, and path planning technologies have been widely applied in indoor and outdoor environments. However, in the dim and narrow environment of phosphate ore mining roadways, the related technologies that rely on robot technology and autonomous navigation technology to complete the autonomous driving of transport vehicles in the roadways, enabling them to drive to the designated working face and cooperate with other mining equipment, are still in the research stage and have not been well applied in practice. Summary of the Invention

[0003] The main purpose of the present invention is to provide an unmanned transport vehicle system and application method applicable to phosphate ore mining roadways based on multi-sensor data fusion. The unmanned driving system of the transport vehicle consists of a perception system, a control system, and an execution system. Through the cooperation of each system, the unmanned transport vehicle can complete autonomous navigation tasks.

[0004] To achieve the above technical features, the object of the present invention is realized as follows: An unmanned transport vehicle system applicable to phosphate ore mining roadways, comprising:

[0005] An execution system, which is used to carry the entire transport vehicle system, and the execution system includes an unmanned transport vehicle;

[0006] A perception system, which relies on sensors to provide environmental information for the unmanned transport vehicle of the execution system and assist the unmanned transport vehicle to complete navigation and positioning;

[0007] A control system, which includes two modules, a decision-making module and a planning module, and uses the environmental information from the perception system to obtain the control signal of the unmanned transport vehicle;

[0008] The execution system receives the control signal issued by the control system and sends execution information such as direction, speed, acceleration, heading angle, and driving trajectory to the unmanned transport vehicle, enabling the unmanned transport vehicle to avoid local obstacles and track the global path, and then complete the unmanned driving task in the phosphate ore mining roadway.

[0009] The perception system includes an LED lighting source mounted on the unmanned transport vehicle. The LED lighting source provides limited-range lighting for the phosphate mining roadway. Then, a depth camera mounted on the unmanned transport vehicle uses the LED light source to capture environmental images of the phosphate mining roadway, obtains depth information from the environmental images, and then performs morphological, clustering, and binarization operations on the environmental images to calculate the centroid of the environmental images, guiding the unmanned transport vehicle to move forward in the correct direction with the centroid. At the same time, a 3D lidar mounted on the unmanned transport vehicle scans the phosphate mining roadway to obtain roadway point clouds. After filtering and region-of-interest segmentation processing of the point clouds, the road surface point clouds are extracted, and then the Alpha Shapes algorithm is used to extract the lane boundaries of the road surface point clouds to constrain the unmanned transport vehicle to drive in the middle of the roadway. When the unmanned transport vehicle is driving in the roadway, laser inertial positioning is adopted. Data collection and processing are carried out in a loosely coupled manner. The 3D lidar performs state estimation, and the IMU data is used as the observation value, and then the Kalman filter is used to fuse the data to complete the task of positioning the unmanned transport vehicle.

[0010] The decision-making module determines the driving speed and mode of the unmanned transport vehicle according to the environmental information. Within the range defined by the decision-making, the planning module combines the vehicle kinematic model, obstacle information in the roadway, driving tasks, and traffic rules to solve for a smooth and continuous local driving trajectory, and sends the speed and trajectory information to the execution system.

[0011] An operation method for an unmanned transport vehicle system applicable to a phosphate mining roadway includes the following steps:

[0012] Step 1: According to the phosphate mining roadway map, plan the global path for the navigation of the unmanned transport vehicle.

[0013] Step 2: Provide limited-range lighting through the LED lighting source mounted on the unmanned transport vehicle.

[0014] Step 3: The depth camera captures environmental images of the phosphate mining roadway environment and performs corresponding processing on the environmental images.

[0015] Step 4: The information of the environmental images is further processed to obtain the centroid of the open area, guiding the unmanned transport vehicle to move forward in the correct direction.

[0016] Step 5: The 3D lidar mounted on the unmanned transport vehicle performs environmental scanning to obtain environmental point cloud data.

[0017] Step 6: Process the point cloud data to obtain the ground point cloud.

[0018] Step 7: Process the ground point cloud and obtain the roadway boundary through the Alpha Shapes algorithm.

[0019] Step 8: Real-time positioning of the unmanned transport vehicle.

[0020] Step 9: The sensor detects obstacles in real time;

[0021] Step 10: Local path planning to avoid obstacles;

[0022] Step 11: The unmanned transport vehicle completes the autonomous navigation task and travels to the designated working face;

[0023] Step 12: The unmanned transport vehicle cooperates with the mucking machine to load and transport the ore slag.

[0024] The specific process of step 1 is as follows: The unmanned transport vehicle, based on the existing environmental map of the phosphate ore mining roadway and combined with the position of the working face to be reached, uses the global path planning algorithm to obtain the optimal path from the starting point to the working face. The unmanned transport vehicle tracks the established path, plans the driving speed, and starts the navigation movement;

[0025] The specific process of step 2 is as follows: The LED lighting source carried by the unmanned transport vehicle provides a light source within a limited range. This lighting source evenly illuminates the phosphate ore mining roadway with the unmanned transport vehicle as the origin. However, in the farther roadway, the lighting will weaken, leaving the farther roadway still in the unlit dark open area;

[0026] The specific process of step 3 is as follows: The depth camera carried by the unmanned transport vehicle takes pictures of the roadway within the lighting range. The content of the pictures is the illuminated roadway and the dark unlit roadway at the end of the light source. First, by processing the pictures, the depth distance is obtained, and morphological processing, image smoothing, and clustering operations are carried out. Then the pictures are binarized to obtain the processed pictures;

[0027] The specific process of step 4 is as follows: The in-vehicle computer of the unmanned transport vehicle uses the binarized pictures obtained in step 3 to calculate the moments of the open areas in the binary results, extract the centroid, use the centroid as the navigation heading point, and the depth information as the target distance, and send them to the actuator of the unmanned transport vehicle for direction control, so that the unmanned transport vehicle tracks the established path and travels in the correct direction.

[0028] The specific process of step 5 is as follows: The 3D lidar carried by the unmanned transport vehicle scans the phosphate ore mining roadway to obtain the point cloud data of the roadway within the sensing range of the 3D lidar;

[0029] The specific process of step 6 is as follows: The point cloud data obtained in step 5 is processed. First, the region of interest segmentation is performed on the point cloud data to reduce the loss of computing memory and improve the processing speed. Then, the point cloud data within the region of interest is separated into the ground point cloud and the point cloud of the phosphate ore mining roadway wall on the ground, and the ground point cloud data is retained;

[0030] The specific process of step 7 is as follows: The on-vehicle computer processes the ground point cloud obtained in step 6, obtains the boundary information of the phosphate ore mining roadway through the AlphaShapes algorithm, calculates the central position of the roadway, and then sends the information to the unmanned transport vehicle controller to restrict the unmanned transport vehicle to drive in the middle of the roadway, avoiding contact or collision with the roadway wall. The processing of the above steps enables the unmanned transport vehicle to autonomously and safely drive along the globally planned path in the phosphate ore mining roadway.

[0031] The specific process of step 8 is as follows: When the unmanned transport vehicle is driving in the phosphate ore mining roadway, it needs to locate its position in the roadway in real time. The unmanned transport vehicle is located by combining a 3D lidar and an IMU. In a loose coupling form, the 3D lidar is used for the state estimation of the unmanned transport vehicle, and the IMU is used as the observation data. The extended Kalman filter is used to fuse the data for accurate, robust, and drift-free long-term attitude estimation to locate the position of the unmanned transport vehicle; Since the linear velocity estimation is inaccurate, the heading information provided by the centroid point is used as a direction filter to accurately estimate the driving speed.

[0032] The specific process of step 9 is as follows: There are static and dynamic obstacles such as staff, other vehicles, and ore blocks in the phosphate ore mining roadway. Different sensors are used to sense these obstacles in real time; The 3D lidar has a relatively long sensing range. In the phosphate ore mining roadway, the 3D lidar is used to detect transport vehicles at a relatively long distance. By comparing adjacent frames, the speed and distance information of oncoming transport vehicles are obtained, leaving enough planning time for the unmanned transport vehicle to plan; Affected by its own structure, the depth camera has a limited detection distance. Therefore, the depth camera is used to capture the near roadway environment information in real time, and then the deep learning algorithm is used to identify and track the objects existing in the environment, and estimate the movement speed and trajectory of dynamic objects.

[0033] The specific process of step 10 is as follows: The phosphate mining roadway serves as a one-way lane. When encountering other vehicles, it is rather troublesome to pass each other. Therefore, refuge chambers are built at certain intervals in the roadway for two oncoming vehicles to pass each other. During the navigation of the driverless transport vehicle, when encountering obstacles, it is necessary to plan a local path for avoidance according to the type of the obstacle and in combination with the traffic rules in the roadway. First, according to the positioning method in step 8, determine its own position in the roadway, and in combination with the roadway map, determine the position of the refuge chamber. At the same time, the on-vehicle computer fuses the picture information to establish a local three-dimensional scene map with semantic information. Then, when the lidar carried by the driverless transport vehicle detects an oncoming vehicle according to the method in step 9, transmit the above information to the decision-making module of the driverless transport vehicle. It decides its own driving speed according to the speed and position of the oncoming vehicle, and the roadway traffic rules of "a lighter vehicle gives way to a heavier vehicle" and "facilitating avoidance", and uses the semantic map to judge the size of the refuge chamber, and determines that the driverless transport vehicle enters the refuge chamber in a semi-entry or full-entry manner for avoidance. When the depth camera detects a staff member according to the method in step 9, the decision-making module determines the width of the roadway at this time in combination with the semantic map according to the speed and trajectory of the pedestrian, and decides to decelerate or stop to avoid the staff member. Finally, after the decision-making module makes a decision, the joint planning module makes a local path plan and then issues it to the execution system.

[0034] The specific process of step 11 is as follows: Through steps 1-10, the driverless transport vehicle completes the global navigation task and relies on local path planning for obstacle avoidance and safely drives to the specified target point;

[0035] The specific process of step 12 is as follows: The depth camera carried by the driverless transport vehicle collects environmental pictures again and transmits them to the pre-trained neural network. By detecting the environmental features of the working face or the mucking machine equipment, judge whether the driverless transport vehicle has reached the mining working face. When it reaches the working face, in combination with the depth information of the environmental picture, adjust the distance between the driverless transport vehicle and the mucking machine to make them cooperate, and then start the mucking machine to load and transport the ore slag.

[0036] The specific process for the driverless transport vehicle to complete autonomous navigation includes the following steps:

[0037] S1 is the start;

[0038] S2 is to start the driverless transport vehicle to make the entire system of the driverless transport vehicle start working;

[0039] S3 is for the driverless transport vehicle to navigate by following the global path plan:

[0040] Step A1: The on-vehicle computer sets the starting point and the ending point according to the existing phosphate mining roadway map, and uses relevant global path planning algorithms to obtain the global optimal path from the starting point to the working face;

[0041] Step A2: The driverless transport vehicle follows the global path and navigates in the phosphate mining roadway;

[0042] S4 is for the depth camera and 3D lidar to sense environmental data:

[0043] Step B1: The driverless transport vehicle moves in the roadway, uses the depth camera and 3D lidar to sense the roadway environment, the lighting source installed on the driverless transport vehicle provides illumination, and the depth camera takes environmental pictures of the roadway environment in real time with the help of the light source;

[0044] Step B2: During the movement of the driverless transport vehicle, the 3D lidar scans the roadway environment in real time to obtain roadway point cloud information;

[0045] S5 is for the 3D lidar and IMU to sense the state of the driverless transport vehicle:

[0046] Step C1: During the movement of the driverless transport vehicle, the two modules act as independent entities and respectively record the state of the driverless transport vehicle; the 3D lidar performs real-time state estimation on the driverless transport vehicle system to obtain the state inside the driverless transport vehicle system; the IMU records the acceleration, speed and heading angle information of the driverless transport vehicle;

[0047] S6 is for the depth camera and 3D lidar data processing:

[0048] Step D1: Using the environmental pictures obtained in Step B1, first obtain the depth information of the pictures, then perform a series of operations on the pictures such as denoising, grayscale conversion, binarization, and clustering, and then obtain the centroid of the open area in the pictures through moment calculation;

[0049] Step D2: Using the roadway point cloud data obtained in B2, first perform point cloud filtering to reduce the number of point clouds, obtain the point clouds of the region of interest through point cloud segmentation, then filter the point clouds of the roadway wall and the wall top in the region of interest, retain the road surface point clouds, and use the Alpha Shapes algorithm to extract the road boundary;

[0050] Step D3: Using the centroid points obtained in D1 and the boundary information of the road in D2, the driverless transport vehicle obtains a safe navigation direction;

[0051] S7 is for the 3D lidar and IMU data processing:

[0052] Step E1: Using the extended Kalman filter for the driverless transport vehicle state estimation information and the IMU observation information in Step C1, obtain the real-time state of the driverless transport vehicle in the roadway and locate the position of the driverless transport vehicle in the roadway;

[0053] S8 is for the navigation and positioning information:

[0054] Step F1: The navigation and positioning information is transmitted through data and sent to the decision planner;

[0055] S9 is the decision-making planner:

[0056] Step G1: The navigation and positioning information is transmitted from Step F1. Combining with the environmental information, the decision-making module is responsible for determining the movement mode of the driverless transport vehicle, including normal driving, slow down, or enter the refuge chamber to stop and avoid; within the range defined by the decision-making module, the planning module combines the vehicle kinematic model, obstacle information in the environment, driving tasks, etc. to solve and obtain a smooth and continuous local driving trajectory;

[0057] S10 means that the driverless transport vehicle has not reached the designated position:

[0058] Step H1: By detecting and identifying the road surface characteristics of the working face, dangerous floating stones on the roadway wall and roof, it is judged that the driverless vehicle has not reached the designated position;

[0059] S11 means that the driverless transport vehicle continues to move forward:

[0060] Step I1: The driverless transport vehicle has not reached the designated position, and it continues to move forward by tracking the global path until it reaches the designated working face;

[0061] S12 means that the driverless transport vehicle has reached the designated position:

[0062] Step J1: By detecting and identifying the characteristics of the working face road surface, mucking machine, roadway wall and roof dangerous floating stones, it is judged that the driverless vehicle has reached the designated position, and the navigation task ends;

[0063] S13 means that the driverless transport vehicle cooperates with other working devices to complete the operation:

[0064] Step K1: Through the pictures taken by the depth camera, identify the mucking machine, obtain the distance to the mucking machine, and then adjust the position of the driverless vehicle to cooperate with the mucking machine to load and transport the ore slag.

[0065] The process of using the driverless transport vehicle system to carry a series of mining equipment to complete relevant unmanned operations includes the following steps:

[0066] S1.1 is the start;

[0067] S1.2 is to use the whole set of solutions of the present invention to complete different tasks by using different mining equipment carried by the driverless transport vehicle on the basis of realizing the driverless of the mining vehicle;

[0068] S1.3 is that the driverless transport vehicle carrying the mining equipment travels along the globally planned path, and judges whether the driverless transport vehicle has reached the designated position by identifying special markers or special scenes on the working face;

[0069] S1.4 is that after the driverless transport vehicle carrying the mining equipment reaches the designated position, it collects the working condition information of the working face through the vehicle-mounted sensors;

[0070] S1.5 processes the environmental information of the roadway working face collected in S1.4. For example, for the point cloud obtained by using lidar to scan the environment, it determines the position, height, size, and direction angle of dangerous loose rocks; for the pictures taken by the depth camera, it fuses the point cloud to reconstruct the three-dimensional scene model of the rock wall to be drilled, and uses it to determine the drilling position and density information of the rock surface;

[0071] S1.6 is coordinate transformation:

[0072] Step A1.1: Using the environmental processing data in S1.5, perform coordinate transformation with the unmanned transport vehicle and on-vehicle mining equipment to obtain the relative position relationship between the working face to be processed and the mining equipment, and then adjust the position of the equipment to process the specified working position and complete related operations;

[0073] S1.7 is unmanned operation:

[0074] Step B1.1: According to the position relationship between the working face and the on-vehicle equipment obtained in Step A1.1, use the mining equipment to accurately complete the unmanned operation, improving work efficiency while reducing operation errors;

[0075] S1.8 is to judge that the relevant unmanned operation is not completed;

[0076] Step C1.1: Judge that the unmanned operation of the current working face has not ended, continue to return to the data processing part, repeat the processing of S1.5 to S1.7 for the data of the next position of the working face, and continue the unmanned operation;

[0077] S1.9 is to judge that the relevant unmanned operation is completed:

[0078] Step D1.1: Complete the current unmanned operation, and the unmanned transport vehicle carrying the relevant mining equipment exits the current working face;

[0079] S1.10 is the end.

[0080] An unmanned transport vehicle system and operation method applicable to phosphate mining roadways proposed by the present invention can achieve the following technical results by adopting the above technical solutions:

[0081] 1. The present invention can achieve autonomous driving in the roadway in the dim environment of the phosphate mining roadway relying on the on-vehicle sensors of the transport vehicle and safely navigate to the specified position.

[0082] 2. The present invention uses the light source with limited external lighting distance, takes pictures of the environment with a depth camera to obtain the depth information of the environment pictures, extracts the centroid of the open area of the phosphate mining roadway, uses the centroid as the heading point to correct the forward direction of the unmanned transport vehicle in the roadway, and follows the global planned path.

[0083] 3. The present invention uses 3D lidar and IMU to achieve real-time positioning of unmanned transport vehicles in phosphate mining roadways. The lidar is used as the state estimation quantity, and the IMU is used as the observation quantity to update the EKF, realizing long-term stable attitude estimation. At the same time, aiming at the inaccurate estimation of the linear velocity of the unmanned transport vehicle, the centroid of the picture is used as the direction filter to accurately estimate the driving speed of the unmanned transport vehicle.

[0084] 4. The present invention uses 3D lidar and depth camera data for fusion mapping, creates a local three-dimensional scene map with semantic information, and determines the size of the refuge chamber and the roadway width through the semantic map, so as to decide the avoidance method of the unmanned transport vehicle.

[0085] 5. The present invention uses deep learning algorithms to identify the mucking machine on the phosphate mining face, and then through the judgment of the state and relative position, completes the precise cooperation between the transport vehicle and the mucking machine, realizing autonomous unmanned ore slag loading and transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] The present invention will be further described below with reference to the drawings and embodiments.

[0087] Figure 1 It is the framework diagram of the unmanned driving system of the present invention.

[0088] Figure 2 It is the method diagram for extracting the centroid of the picture of the present invention.

[0089] Figure 3 It is the method diagram for extracting the roadway boundary of the present invention.

[0090] Figure 4 It is the positioning method diagram of the unmanned transport vehicle of the present invention.

[0091] Figure 5 It is the local three-dimensional scene reconstruction method diagram of the present invention.

[0092] Figure 6 It is the passing strategy process diagram of the unmanned transport vehicle of the present invention.

[0093] Figure 7 It is the autonomous navigation implementation process diagram of the unmanned transport vehicle of the present invention.

[0094] Figure 8 It is the unmanned operation process diagram of various unmanned mining equipment using the solution of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0095] Example 1:

[0096] See Figure 1-8, the main object of the present invention is to provide an unmanned transport vehicle based on multi-sensor data fusion and applicable to phosphate mining roadways. The unmanned driving system of the transport vehicle consists of a perception system, a control system and an execution system. Through the cooperation of each system, the unmanned transport vehicle can complete autonomous navigation tasks. First, the on-vehicle computer plans a global path according to the map of the phosphate mining roadway and drives safely along the established route. Then, the unmanned transport vehicle uses the fusion data of the on-vehicle multi-sensors to perceive the surrounding environment of the vehicle and establish an accurate environmental model; by obtaining information on the position of the vehicle itself and dynamic and static obstacles such as other vehicles, people, and ores, and based on the driving rules of the roadway, it plans the local driving route of the vehicle. Finally, the unmanned transport vehicle drives along the autonomously planned route to complete the task of navigating to the target working face; and then through the perception and recognition of the sensors, it obtains the cooperation information with other mining equipment, adjusts the distance, and cooperates with each other to complete the work.

[0097] To achieve the above object, the technical solution of the present invention is as follows: 1. The on-vehicle computer combines the starting point of the unmanned transport vehicle and the position of the working face to be reached, uses the global path planning algorithm to plan the optimal path on the roadway map, and sends it to the execution system of the unmanned transport vehicle, so that the unmanned transport vehicle travels along the established route. To prevent deviation from the established path due to control and other problems, a method of combining sensor data to guide the unmanned transport vehicle to travel in the correct direction is proposed: use the depth camera mounted on the unmanned transport vehicle to take pictures of the roadway, and according to the depth pictures, extract the centroid of the open area of the roadway with depth information, and use the centroid as the direction to guide the unmanned transport vehicle to move forward in the correct direction in the phosphate mining roadway; at the same time, in order to prevent the unmanned transport vehicle from hitting the wall when turning due to its large body width during traveling along the established route, a roadway boundary detection method is proposed: use the 3D lidar mounted on the unmanned transport vehicle to scan the roadway to obtain the roadway point cloud data. Then extract the road boundary from the roadway point cloud data, and use the boundary of the road to restrict the unmanned transport vehicle to the middle of the road in the roadway and keep moving in the correct direction. 2. During the driving process of the unmanned transport vehicle, when the sensor detects a mine car, pedestrian or ore, it is necessary to combine the positioning to determine its own position in the roadway, and take measures such as entering the refuge chamber to avoid, stopping, slowing down, and normal driving to meet or give way. First, the present invention adopts the method of combining 3D lidar and IMU for positioning to complete the real-time positioning task in the phosphate mining roadway; then, as the unmanned transport vehicle travels, sensors such as on-vehicle lidar and depth camera collect environmental data in the roadway in real time, and through the perception data of the sensors, identify the type of obstacles, judge the moving speed and distance; finally, after the positioning is completed, combined with the environmental information, fuse the roadway point cloud data and image information to perform local three-dimensional scene reconstruction to obtain a three-dimensional scene map with semantic information; the decision-making module of the unmanned transport vehicle decides the obstacle avoidance method of the unmanned transport vehicle according to the semantic map; then jointly plan the module to perform local trajectory planning and send it to the execution system of the transport vehicle. After avoiding the obstacles, resume the original driving state and continue to drive. 3. After the unmanned transport vehicle completes the navigation task according to the established route, it judges whether the unmanned transport vehicle has traveled to the designated roadway working face by identifying environmental features or equipment such as mucking machines. Then adjust its own position, cooperate with the mucking machine, and start the mucking machine to load and transport the ore slag. 4. The present invention realizes the driverless of the transport vehicle in the phosphate mining roadway through the above solutions, and cooperates with the mucking machine to complete the loading and transportation of the ore slag. The whole set of solutions can also be used to design the driverless of a series of mining equipment. For example, use a driverless vehicle to carry a scaling equipment to form an unmanned scaling vehicle. Wait for it to autonomously drive to the designated working face, identify dangerous loose rocks through sensors, and combine coordinate transformation to obtain the relevant position information of the scaling equipment and dangerous loose rocks, and accurately complete the unmanned scaling operation; use a driverless vehicle to carry a drilling equipment to form an unmanned drilling vehicle. Wait for it to autonomously drive to the designated working face, combine sensors to identify the area of the rock wall, and reasonably arrange the number and position of drill holes.

[0098] Example 2:

[0099] Refer to Figure 1-6 , an unmanned transport vehicle system applicable to phosphate mining roadways, which includes a sensing system 101, a control system 102, and an execution system 103. The sensing system 101 relies on sensors to provide environmental information for the unmanned transport vehicle and assist the unmanned transport vehicle to complete navigation and positioning. First, the LED lighting source mounted on the unmanned transport vehicle illuminates a limited range of the phosphate mining roadway. Then, the depth camera 104 mounted on the unmanned transport vehicle uses the LED light source to take environmental pictures of the phosphate mining roadway, obtains depth information from the pictures, and then performs operations such as morphology, clustering, and binarization on the pictures to calculate the centroid of the pictures, and guides the unmanned transport vehicle to move in the correct direction with the centroid. At the same time, the 3D lidar 105 mounted on the unmanned transport vehicle scans the phosphate mining roadway to obtain roadway point clouds. After filtering and region of interest segmentation processing of the point clouds, the road surface point clouds are extracted, and then the Alpha Shapes algorithm is used to extract the boundaries of the road surface point clouds to constrain the unmanned transport vehicle to drive in the middle of the roadway. When the unmanned transport vehicle is driving in the roadway, laser inertial navigation positioning is adopted. Data collection and processing are carried out in a loosely coupled manner. The lidar 105 performs state estimation, and the IMU106 data is used as an observation value, and then the Kalman filter is used to fuse the data to complete the task of positioning the unmanned transport vehicle. The control system 102 includes two modules, a decision-making module 107 and a module planning 108, which uses the environmental information from the sensing system to obtain the control signal of the unmanned transport vehicle. First, the decision-making module 107 determines the driving speed and mode of the unmanned transport vehicle according to the environmental information. Within the range defined by the decision, the planning module 108 combines the vehicle kinematic model, obstacle information in the roadway, driving tasks, and traffic rules, etc., to solve for a smooth and continuous local driving path, and sends the speed and trajectory information to the execution system 103. The execution system 103 receives the control signal sent by the control system, sends speed, acceleration, and driving trajectory execution information to the unmanned transport vehicle, enables the unmanned transport vehicle to avoid local obstacles, and tracks the global path to complete the navigation task.

[0100] Example 3:

[0101] An operation method of an unmanned transport vehicle system applicable to phosphate mining roadways includes the following steps:

[0102] Step 1: According to the phosphate mining roadway map, plan the global path for the unmanned transport vehicle to navigate; the unmanned transport vehicle, based on the existing phosphate mining roadway environment map, combines the position of the working face to be reached, and uses the global path planning algorithm to obtain the best path from the starting point to the working face. The unmanned transport vehicle tracks the established path, plans the driving speed, and starts the navigation movement;

[0103] Step 2: Provide limited-range illumination through the LED lighting source carried by the driverless transport vehicle. The LED lighting source carried by the driverless transport vehicle provides a limited-range light source, which evenly illuminates the phosphate mining roadway with the driverless transport vehicle as the origin. However, in the roadway at a greater distance, the illumination will weaken, leaving the roadway in the distance still in the unlit dark open area.

[0104] Step 3: The depth camera 104 captures environmental pictures of the phosphate mining roadway environment and performs corresponding processing on the environmental pictures. The depth camera 104 carried by the driverless transport vehicle captures pictures of the roadway within the illumination range, which show the illuminated roadway and the dark unlit roadway at the end of the light source. First, the depth distance is obtained through picture processing, followed by morphological processing, image smoothing, and clustering operations. Then, the picture is binarized to obtain the processed picture.

[0105] Step 4: The information of the environmental pictures is further processed to obtain the centroid of the open area, guiding the driverless transport vehicle to move in the correct direction. The in-vehicle computer of the driverless transport vehicle uses the binarized picture obtained in Step 3 to calculate the moments of the open area in the binary result, extract the centroid, use the centroid as the navigation heading point, and the depth information as the target distance, and send them to the actuator of the driverless transport vehicle for direction control, enabling the driverless transport vehicle to track the predefined path and drive in the correct direction.

[0106] Step 5: The 3D lidar 105 carried by the driverless transport vehicle performs environmental scanning to obtain environmental point cloud data. The 3D lidar 105 carried by the driverless transport vehicle scans the phosphate mining roadway to obtain the point cloud data of the roadway within the sensing range of the 3D lidar.

[0107] Step 6: Process the point cloud data to obtain the ground point cloud. Process the point cloud data obtained in Step 5. First, perform region-of-interest segmentation on the point cloud data to reduce the loss of computing memory and improve the processing speed. Then, separate the point cloud data within the region of interest into the ground point cloud and the point cloud of the phosphate mining roadway wall on the ground, and retain the ground point cloud data.

[0108] Step 7: Process the ground point cloud to obtain the roadway boundary through the Alpha Shapes algorithm. The in-vehicle computer processes the ground point cloud obtained in Step 6, obtains the boundary information of the phosphate mining roadway through the Alpha Shapes algorithm, calculates the central position of the roadway, and then sends the information to the driverless transport vehicle controller to restrict the driverless transport vehicle to drive in the middle of the roadway, avoiding contact or collision with the roadway wall. The processing of the above steps enables the driverless transport vehicle to drive autonomously and safely in the phosphate mining roadway according to the global planned path.

[0109] Step 8: Real-time positioning of the driverless transport vehicle; when the driverless transport vehicle is traveling in the phosphate ore mining roadway, it needs to locate its position in the roadway in real time. The 3D lidar 105 and IMU 106 are combined for the positioning of the driverless transport vehicle. In a loose coupling form, the 3D lidar 105 is used for the state estimation of the driverless transport vehicle, and the IMU 106 is used as the observed data. The extended Kalman filter is used to fuse the data for accurate, robust, and drift-free long-term attitude estimation to locate the position of the driverless transport vehicle; due to inaccurate linear velocity estimation, the heading information provided by the centroid point is used as a direction filter to accurately estimate the driving speed; here, the IMU 106 is an inertial measurement unit.

[0110] Step 9: Sensors detect obstacles in real time; there are dynamic and static obstacles such as staff, other vehicles, and ore blocks in the phosphate ore mining roadway. Different sensors are used to perceive these obstacles in real time; the 3D lidar 105 has a relatively long sensing range. In the phosphate ore mining roadway, the 3D lidar 105 is used to detect transport vehicles at a relatively long distance. By comparing adjacent frames, the speed and distance information of oncoming transport vehicles are obtained, leaving enough planning time for the driverless transport vehicle to plan; affected by its own structure, the depth camera 104 has a limited detection distance. Therefore, the depth camera is used to capture the near roadway environment information in real time, and then the deep learning algorithm is used to identify and track the objects existing in the environment, and estimate the movement speed and trajectory of dynamic objects.

[0111] Step 10: Local path planning to avoid obstacles; the phosphate ore mining roadway is a single-lane passage. When encountering other vehicles, it is rather troublesome to pass each other. Therefore, refuge chambers are built at certain intervals in the roadway for two oncoming vehicles to pass each other; when the driverless transport vehicle is navigating forward and encounters an obstacle, it needs to plan a local path for avoidance according to the type of the obstacle and in combination with the traffic rules in the roadway; first, according to the positioning method in Step 8, determine its own position in the roadway, and in combination with the roadway map, determine the position of the refuge chamber. At the same time, the on-vehicle computer fuses the picture information to establish a local three-dimensional scene map with semantic information; then, when the lidar carried by the driverless transport vehicle detects an oncoming vehicle according to the method in Step 9, the above information is transmitted to the decision-making module 107 of the driverless transport vehicle. It decides its own driving speed according to the speed and position of the oncoming vehicle and the roadway traffic rules of "a lighter vehicle gives way to a heavier vehicle" and "facilitating avoidance", and uses the semantic map to judge the size of the refuge chamber to determine whether the driverless transport vehicle enters the refuge chamber in a semi-entry or full-entry manner for avoidance; when the depth camera 104 detects staff according to the method in Step 9, the decision-making module 107 judges the width of the roadway at this time in combination with the semantic map according to the speed and trajectory of the pedestrian, and decides to decelerate or stop to avoid the staff; finally, after the decision-making module 107 makes a decision, the joint planning module 108 makes a local path plan and then issues it to the execution system 103.

[0112] Step 11: The driverless transport vehicle completes the autonomous navigation task and travels to the designated working face; through Steps 1 - 10, the driverless transport vehicle completes the global navigation task, relies on local path planning for obstacle avoidance, and safely travels to the designated target point;

[0113] Step 12: The driverless transport vehicle cooperates with the mucking machine to load and transport ore slag. The depth camera 104 carried by the driverless transport vehicle collects environmental pictures again, transmits them to the pre-trained neural network, and determines whether the driverless transport vehicle has reached the mining working face by detecting the environmental features of the working face or the equipment of the mucking machine; when reaching the working face, combined with the depth information of the environmental pictures, adjust the distance between the driverless transport vehicle and the mucking machine to make the two cooperate, and then start the mucking machine to load and transport ore slag.

[0114] Example 4:

[0115] See Figure 7 , the specific process for the driverless transport vehicle to complete autonomous navigation includes the following steps:

[0116] S1 is the start;

[0117] S2 is to start the driverless transport vehicle so that the entire system of the driverless transport vehicle starts to work;

[0118] S3 is for the driverless transport vehicle to navigate by following the global path planning:

[0119] Step A1: The on-vehicle computer sets the starting point and the ending point according to the existing map of the phosphate ore mining roadway, and uses relevant global path planning algorithms to obtain the global optimal path from the starting point to the working face;

[0120] Step A2: The driverless transport vehicle follows the global path and navigates in the phosphate ore mining roadway;

[0121] S4 is for the depth camera and 3D lidar to sense environmental data:

[0122] Step B1: The driverless transport vehicle moves in the roadway, uses the depth camera and 3D lidar to sense the roadway environment, the lighting source installed on the driverless transport vehicle provides lighting, and the depth camera uses the light source to take environmental pictures of the roadway environment in real time;

[0123] Step B2: During the driving process of the driverless transport vehicle, the 3D lidar scans the roadway environment in real time to obtain the roadway point cloud information;

[0124] S5 is for the 3D lidar and IMU to sense the state of the driverless transport vehicle:

[0125] Step C1: During the movement of the driverless transport vehicle, the two modules act as independent entities and respectively record the status of the driverless transport vehicle; the 3D lidar performs real-time state estimation on the driverless transport vehicle system to obtain the internal state of the driverless transport vehicle system; the IMU records the acceleration, speed and heading angle information of the driverless transport vehicle.

[0126] S6 is the data processing of the depth camera and the 3D lidar:

[0127] Step D1: Using the environmental image obtained in Step B1, first obtain the depth information of the image, then perform a series of operations on the image such as denoising, grayscaling, binarization, and clustering, and then obtain the centroid of the open area in the image through moment calculation.

[0128] Step D2: Using the roadway point cloud data obtained in B2, first perform point cloud filtering to reduce the number of point clouds, obtain the point clouds of the region of interest through point cloud segmentation, then filter the point clouds of the roadway wall and the wall top in the region of interest, retain the pavement point clouds, and use the Alpha Shapes algorithm to extract the road boundary.

[0129] Step D3: Using the centroid point obtained in D1 and the boundary information of the road in D2, the driverless transport vehicle obtains a safe navigation direction.

[0130] S7 is the data processing of the 3D lidar and the IMU:

[0131] Step E1: Using the extended Kalman filter for the driverless transport vehicle state estimation information and the observation information of the IMU in Step C1, obtain the real-time state of the driverless transport vehicle in the roadway and locate the position of the driverless transport vehicle in the roadway.

[0132] S8 is the navigation and positioning information:

[0133] Step F1: The navigation and positioning information is transmitted through data and sent to the decision planner.

[0134] S9 is the decision planner:

[0135] Step G1: The navigation and positioning information transmitted in Step F1, combined with the environmental information, the decision module is responsible for determining the movement mode of the driverless transport vehicle, normal driving, slow down or enter the refuge chamber to stop and avoid; within the range defined by the decision module, the planning module combines the vehicle kinematic model, obstacle information in the environment, driving tasks, etc. to solve for a smooth and continuous local driving trajectory.

[0136] S10 is that the driverless transport vehicle has not reached the designated position:

[0137] Step H1: Determine that the driverless vehicle has not reached the designated position by detecting and identifying the pavement characteristics of the working face, dangerous floating stones on the roadway wall and the roof of the roadway.

[0138] S11 is for the driverless transport vehicle to continue moving forward:

[0139] Step I1: When the driverless transport vehicle has not reached the designated position, it continues to move forward following the global path until it reaches the designated working face.

[0140] S12 is for the driverless transport vehicle to reach the designated position:

[0141] Step J1: By detecting and identifying the characteristics of the working face road surface, mucking loader, roadway wall, and dangerous floating rocks on the roadway roof, it is determined that the driverless vehicle has reached the designated position, and the navigation task ends.

[0142] S13 is for the driverless transport vehicle to cooperate with other working devices to complete the operation:

[0143] Step K1: Through the pictures taken by the depth camera, identify the mucking loader, obtain the distance to the mucking loader, and then adjust the position of the driverless vehicle to cooperate with the mucking loader to load and transport ore slag.

[0144] Embodiment 5:

[0145] See Figure 8 , the process of using a driverless transport vehicle system to carry a series of mining equipment to complete relevant unmanned operations includes the following steps:

[0146] S1.1 is the start;

[0147] S1.2 is to use the whole set of solutions of the present invention. On the basis of realizing the driverless operation of the mining vehicle, different mining equipment carried by the driverless transport vehicle is used to complete different tasks.

[0148] S1.3 is for the driverless transport vehicle carrying mining equipment to travel along the globally planned path, and by identifying special markers or special scenes on the working face, it is determined whether the driverless transport vehicle has reached the designated position.

[0149] S1.4 is for the driverless transport vehicle carrying mining equipment to collect the working condition information of the working face after reaching the designated position through on-vehicle sensors.

[0150] S1.5 is to process the environmental information of the roadway working face collected in S1.4. For example, using the point cloud obtained by laser radar scanning the environment to determine the position, height, size, and direction angle of dangerous floating rocks; using the pictures taken by the depth camera, and then fusing the point cloud to reconstruct the three-dimensional scene model of the rock wall to be drilled, and using it to determine the drilling position and density information of the rock surface.

[0151] S1.6 is coordinate transformation:

[0152] Step A1.1: Use the environmental processing data in S1.5, perform coordinate transformation with the unmanned transport vehicle and the on-vehicle mining equipment to obtain the relative position relationship between the working face to be processed and the mining equipment, then adjust the position of the equipment to process the specified working position and complete the relevant operations;

[0153] S1.7 is for unmanned operation:

[0154] Step B1.1: According to the position relationship between the working face and the on-vehicle equipment obtained in Step A1.1, use the mining equipment to accurately complete the unmanned operation, improving work efficiency while reducing operation errors;

[0155] S1.8 is to judge that the relevant unmanned operation has not been completed;

[0156] Step C1.1: Judge that the unmanned operation on the current working face has not ended, continue to return to the data processing part, repeat the processing of S1.5 to S1.7 for the data of the next position of the working face, and continue the unmanned operation;

[0157] S1.9 is to judge that the relevant unmanned operation has been completed:

[0158] Step D1.1: Complete the current unmanned operation, and the unmanned transport vehicle carrying the relevant mining equipment exits the current working face;

[0159] S1.10 is the end.

Claims

1. Operating method of an unmanned transport vehicle system applicable to phosphate mining roadways. The unmanned transport vehicle system applicable to phosphate mining roadways includes: An execution system (103) for carrying the entire transport vehicle system. The execution system (103) includes an unmanned transport vehicle. A perception system (101) that relies on sensors to provide environmental information for the unmanned transport vehicle of the execution system (103) and assist the unmanned transport vehicle in completing navigation and positioning. A control system (102) that contains two modules, a decision-making module (107) and a planning module (108), and uses the environmental information from the perception system (101) to obtain control signals for the unmanned transport vehicle. The execution system (103) receives the control signals issued by the control system (102) and sends execution information such as direction, speed, acceleration, heading angle, and driving trajectory to the unmanned transport vehicle, enabling the unmanned transport vehicle to avoid local obstacles and track the global path, and thus complete the unmanned driving task in the phosphate mining roadway. The perception system (101) includes an LED lighting source mounted on the unmanned transport vehicle. The LED lighting source illuminates a limited range of the phosphate mining roadway. Then, a depth camera (104) mounted on the unmanned transport vehicle uses the LED light source to take environmental pictures of the phosphate mining roadway, obtains depth information from the environmental pictures, and then performs morphological, clustering, and binarization operations on the environmental pictures to calculate the centroid of the environmental pictures, guiding the unmanned transport vehicle to move forward in the correct direction with the centroid. At the same time, a 3D lidar (105) mounted on the unmanned transport vehicle scans the phosphate mining roadway to obtain roadway point clouds. After filtering and region-of-interest segmentation processing of the point clouds, the road surface point clouds are extracted, and then the Alpha Shapes algorithm is used to extract the lane boundaries of the road surface point clouds to constrain the unmanned transport vehicle to drive in the middle of the roadway. When the unmanned transport vehicle is driving in the roadway, laser inertial navigation positioning is adopted. Data collection and processing are carried out in a loosely coupled manner. The 3D lidar (105) performs state estimation, and the IMU (106) data is used as an observation value, and then a Kalman filter is used to fuse the data to complete the task of positioning the unmanned transport vehicle. The decision-making module (107) determines the driving speed and mode of the unmanned transport vehicle according to the environmental information. Within the range defined by the decision-making, the planning module (108) combines the vehicle kinematic model, obstacle information in the roadway, driving tasks, and traffic rules to solve for a smooth and continuous local driving trajectory and sends the speed and trajectory information to the execution system (103). It is characterized in that the operating method includes the following steps: Step 1: Plan the global path for the navigation of the unmanned transport vehicle according to the phosphate mining roadway map. Step 2: Provide limited-range lighting through the LED lighting source mounted on the unmanned transport vehicle. Step 3: The depth camera (104) takes environmental pictures of the phosphate mining roadway environment and performs corresponding processing on the environmental pictures. Step 4: The information of the environmental image is further processed to obtain the centroid of the open area, guiding the driverless transport vehicle to move forward in the correct direction; Step 5: The 3D lidar (105) carried by the driverless transport vehicle scans the environment to obtain environmental point cloud data; Step 6: Process the point cloud data to obtain the ground point cloud; Step 7: Process the ground point cloud and obtain the roadway boundary through the Alpha Shapes algorithm; Step 8: Real-time positioning of the driverless transport vehicle; Step 9: The sensor detects obstacles in real time; Step 10: Local path planning to avoid obstacles; Step 11: The driverless transport vehicle completes the autonomous navigation task and travels to the designated working face; Step 12: The driverless transport vehicle cooperates with the mucking machine to load and transport ore; The specific process of the said Step 8 is as follows: When the driverless transport vehicle travels in the phosphate ore mining roadway, it is necessary to locate its position in the roadway in real time. The driverless transport vehicle is located by combining the 3D lidar (105) and the IMU (106). In a loose coupling form, the 3D lidar (105) is used for the state estimation of the driverless transport vehicle, and the IMU (106) is used as the observation data. The extended Kalman filter is used to fuse the data for accurate, robust, and drift-free long-term attitude estimation to locate the position of the driverless transport vehicle. Due to inaccurate linear velocity estimation, the heading information provided by the centroid point is used as a direction filter to accurately estimate the driving speed.

2. The operation method of the driverless transport vehicle system applicable to the phosphate ore mining roadway according to claim 1, characterized in that, The specific process of the said Step 1 is as follows: The driverless transport vehicle, according to the existing environmental map of the phosphate ore mining roadway and in combination with the position of the working face to be reached, uses the global path planning algorithm to obtain the best path from the starting point to the working face. The driverless transport vehicle tracks the established path, plans the driving speed, and starts the navigation movement; The specific process of the said Step 2 is as follows: The LED lighting source carried by the driverless transport vehicle provides a light source within a limited range. This lighting source evenly illuminates the phosphate ore mining roadway with the driverless transport vehicle as the origin. However, in the farther roadway, the lighting will weaken, leaving the farther roadway still in the dark open area that is not illuminated; The specific process of the said Step 3 is as follows: The depth camera (104) carried by the driverless transport vehicle takes pictures of the roadway within the lighting range. The content is the illuminated roadway and the dark unilluminated roadway at the end of the light source. First, the depth distance is obtained through the processing of the picture, and morphological processing, image smoothing, and clustering operations are carried out. Then the picture is binarized to obtain the processed picture; The specific process of the said Step 4 is as follows: The in-vehicle computer of the driverless transport vehicle uses the binarized picture obtained in Step 3 to calculate the moments of the open area in the binarized result, extract the centroid, use the centroid as the heading point for navigation, and the depth information as the target distance, and send them to the actuator of the driverless transport vehicle for direction control, so that the driverless transport vehicle tracks the established path and travels in the correct direction.

3. The operation method of the driverless transport vehicle system applicable to the phosphate rock mining roadway according to claim 1, characterized in that, The specific process of the said Step 5 is as follows: The 3D lidar (105) carried by the driverless transport vehicle scans the phosphate ore mining roadway to obtain the point cloud data of the roadway within the sensing range of the 3D lidar; The specific process of step 6 is as follows: Process the point cloud data obtained in step 5. First, perform region of interest segmentation on the point cloud data to reduce the loss of computing memory and improve the processing speed. Then, separate the point cloud data within the region of interest into the ground point cloud and the point cloud of the phosphate mining roadway wall on the ground, and retain the ground point cloud data. The specific process of step 7 is as follows: The on-vehicle computer processes the ground point cloud obtained in step 6, obtains the boundary information of the phosphate mining roadway through the Alpha Shapes algorithm, calculates the central position of the roadway, and then sends the information to the unmanned transport vehicle controller to restrict the unmanned transport vehicle to drive in the middle of the roadway to avoid contacting or colliding with the roadway wall. The processing of the above steps enables the unmanned transport vehicle to autonomously and safely drive along the global planned path in the phosphate mining roadway.

4. The operating method of the driverless transport vehicle system applicable to the phosphate ore mining roadway according to claim 1, characterized in that, The specific process of step 9 is as follows: There are static and dynamic obstacles such as staff, other vehicles, and ore blocks in the phosphate mining roadway. Different sensors are used to perceive these obstacles in real time. The 3D lidar (105) has a relatively far sensing range. In the phosphate mining roadway, the 3D lidar (105) is used to detect transport vehicles at a relatively long distance. By comparing adjacent frames, the speed and distance information of oncoming transport vehicles are obtained, leaving enough planning time for the unmanned transport vehicle to plan. Affected by its own structure, the depth camera (104) has a limited detection distance. Therefore, the depth camera is used to capture the near roadway environment information in real time, and then the deep learning algorithm is used to identify and track the objects existing in the environment, and estimate the movement speed and trajectory of dynamic objects. The specific process of step 10 is as follows: The phosphate mining roadway is a single-lane passage. When encountering other vehicles, it is rather troublesome to pass each other. Therefore, refuge chambers are built at certain intervals in the roadway for two oncoming vehicles to pass each other. When the unmanned transport vehicle navigates forward and encounters an obstacle, it is necessary to plan a local path for avoidance according to the type of the obstacle and in combination with the traffic rules in the roadway. First, according to the positioning method in step 8, determine its own position in the roadway, and in combination with the roadway map, determine the position of the refuge chamber. At the same time, the on-vehicle computer fuses the picture information to establish a local three-dimensional scene map with semantic information. Then, when the lidar carried by the unmanned transport vehicle detects an oncoming vehicle according to the method in step 9, the above information is transmitted to the decision-making module (107) of the unmanned transport vehicle. It decides its own driving speed according to the speed and position of the oncoming vehicle and the roadway traffic rules of "light vehicle gives way to heavy vehicle" and "easy to avoid", and uses the semantic map to judge the size of the refuge chamber to determine whether the unmanned transport vehicle enters the refuge chamber in a semi-entry or full-entry manner for avoidance. When the depth camera (104) detects a staff member according to the method in step 9, the decision-making module (107) decides to decelerate or stop according to the pedestrian speed and trajectory, and in combination with the semantic map to judge the width of the roadway at this time to avoid the staff member. Finally, after the decision-making module (107) makes a decision, the joint planning module (108) makes a local path plan and then sends it to the execution system (103).

5. The operating method of the driverless transport vehicle system applicable to phosphate mining roadways according to claim 1, characterized in that, The specific process of step 11 is as follows: Through steps 1 - 10, the driverless transport vehicle completes the global navigation task, relies on local path planning for obstacle avoidance, and safely drives to the specified target point; The specific process of step 12 is as follows: The depth camera (104) carried by the driverless transport vehicle collects environmental pictures again, and inputs them into the pre-trained neural network. By detecting the environmental features of the working face or the mucking machine equipment, it is judged whether the driverless transport vehicle has reached the mining working face; When reaching the working face, combined with the depth information of the environmental pictures, the distance between the driverless transport vehicle and the mucking machine is adjusted to make the two cooperate, and then the mucking machine is started to load and transport the ore slag.

6. The operating method of the driverless transport vehicle system applicable to the phosphate ore mining roadway according to claim 1, characterized in that, The specific process for the driverless transport vehicle to complete autonomous navigation includes the following steps: S1 is the start; S2 is to start the driverless transport vehicle, so that the entire system of the driverless transport vehicle starts to work; S3 is for the driverless transport vehicle to navigate following the global path planning: Step A1: The on-vehicle computer sets the starting point and the ending point according to the existing map of the phosphate ore mining roadway, and uses relevant global path planning algorithms to obtain the global optimal path from the starting point to the working face; Step A2: The driverless transport vehicle follows the global path and navigates in the phosphate ore mining roadway; S4 is for the depth camera and 3D lidar to sense environmental data: Step B1: The driverless transport vehicle moves in the roadway, uses the depth camera and 3D lidar to sense the roadway environment, the lighting source installed on the driverless transport vehicle provides lighting, and the depth camera uses the light source to take environmental pictures of the roadway environment in real time; Step B2: During the driving process of the driverless transport vehicle, the 3D lidar scans the roadway environment in real time to obtain the roadway point cloud information; S5 is for the 3D lidar and IMU to sense the state of the driverless transport vehicle: Step C1: During the driving process of the driverless transport vehicle, the two modules, as independent entities, respectively record the state of the driverless transport vehicle; The 3D lidar performs real-time state estimation on the driverless transport vehicle system to obtain the state inside the driverless transport vehicle system; The IMU records the acceleration, speed, and heading angle information of the driverless transport vehicle; S6 is for data processing of the depth camera and 3D lidar: Step D1: Using the environmental pictures obtained in step B1, first obtain the depth information of the pictures, then perform a series of operations on the pictures such as noise reduction, grayscale conversion, binarization, and clustering, and then obtain the centroid of the open area in the pictures through moment calculation; Step D2: Using the roadway point cloud data obtained in B2, first perform point cloud filtering to reduce the number of point clouds, obtain the point clouds of the region of interest through point cloud segmentation, then filter the point clouds of the roadway wall and the wall top in the region of interest, retain the road surface point clouds, and use the Alpha Shapes algorithm to extract the road boundary; Step D3: Using the centroid points obtained in D1 and the road boundary information in D2, the driverless transport vehicle obtains a safe navigation direction; S7 is for data processing of the 3D lidar and IMU: Step E1: Using the extended Kalman filter for the unmanned transport vehicle state estimation information and the IMU observation information in step C1, obtain the real-time state of the unmanned transport vehicle in the roadway and locate the position of the unmanned transport vehicle in the roadway; S8 is for navigation and positioning information: Step F1: The navigation and positioning information is transmitted through data and sent to the decision-making planner. S9 is the decision-making planner: Step G1: The navigation and positioning information transmitted in Step F1 is combined with the environmental information. The decision-making module is responsible for determining the movement mode of the unmanned transport vehicle, i.e., normal driving, slow down or enter the refuge chamber to stop and avoid. Within the range defined by the decision-making module, the planning module solves for a smooth and continuous local driving trajectory by combining the vehicle kinematic model, obstacle information in the environment, driving tasks, etc. S10 means the unmanned transport vehicle has not reached the designated position: Step H1: It is judged that the unmanned vehicle has not reached the designated position by detecting and identifying the road surface characteristics of the working face, dangerous floating rocks on the roadway wall and roof. S11 means the unmanned transport vehicle continues to move forward: Step I1: Since the unmanned transport vehicle has not reached the designated position, it continues to move forward along the global path until it reaches the designated working face. S12 means the unmanned transport vehicle has reached the designated position: Step J1: It is judged that the unmanned vehicle has reached the designated position by detecting and identifying the characteristics of the working face road surface, mucking loader, roadway wall and roof dangerous floating rocks, and the navigation task ends. S13 means the unmanned transport vehicle cooperates with other working devices to complete the operation: Step K1: Through the pictures taken by the depth camera, the mucking loader is identified, and the distance to the mucking loader is obtained. Then the position of the unmanned vehicle is adjusted to cooperate with the mucking loader to load and transport the ore slag.

7. The operation method of the driverless transport vehicle system applicable to the phosphate ore mining roadway according to claim 1, characterized in that The process of using the unmanned transport vehicle system to carry a series of mining equipment to complete relevant unmanned operations includes the following steps: S1.1 is the start; S1.2: Using the whole set of solutions of the present invention, on the basis of realizing the unmanned driving of the mining vehicle, different mining equipment carried by the unmanned transport vehicle is used to complete different tasks. S1.3: The unmanned transport vehicle carrying the mining equipment travels along the globally planned path, and judges whether the unmanned transport vehicle has reached the designated position by identifying special markers or special scenes on the working face. S1.4: After the unmanned transport vehicle carrying the mining equipment reaches the designated position, the working conditions information of the working face is collected by on-vehicle sensors. S1.5: Process the environmental information of the roadway working face collected in S1.

4. For example, use the point cloud obtained by laser radar scanning the environment to determine the position, height, size and direction angle of dangerous floating rocks; use the pictures taken by the depth camera, and then fuse the point cloud to reconstruct the three-dimensional scene model of the rock wall to be drilled, and use it to determine the drilling position and density information of the rock surface. S1.6 is coordinate transformation: Step A1.1: Using the environmental processing data in S1.5, perform coordinate transformation with the unmanned transport vehicle and the on-vehicle mining equipment to obtain the relative position relationship between the working face to be processed and the mining equipment, and then adjust the position of the equipment to process the designated working position and complete the relevant operation. S1.7 is unmanned operation: Step B1.1: According to the position relationship between the working face and the on-vehicle equipment obtained in Step A1.1, use the mining equipment to accurately complete the unmanned operation, improving the work efficiency and reducing operation errors at the same time. S1.8 is to judge that the relevant unmanned operation has not been completed; Step C1.1: Determine that the unmanned operation on the current working face has not ended. Continue to return to the data processing section, repeat the processing from S1.5 to S1.7 for the data at the next position on the working face, and continue the unmanned operation; S1.9 is to determine that the relevant unmanned operation is completed: Step D1.1: Complete the current unmanned operation, and the unmanned transport vehicle carrying the relevant mining equipment exits the current working face; S1.10 is the end.

Images