Coal mine underground transportation robot and control system

Abstract

The present invention relates to the technical field of transport robots, and in particular to a coal mine underground transport robot and control system, comprising a transport trolley and a carriage, wherein the carriage is bolted to the upper surface of the transport trolley, a base is installed on the transport trolley at one side of the carriage, an obstacle clearing component is arranged on the base, and the obstacle clearing component is used to clear obstacles on the coal mine road, and a bracket is also installed on the base, and a depth camera, an image analysis module and a controller are installed on the bracket. The coal mine underground transport robot and control system provided by the present invention are equipped with an obstacle clearing component and a depth camera, and can collect images of obstacles on the transport path, and obtain the state coefficient of the obstacle based on image analysis, and then judge whether the obstacle can be cleared by the obstacle clearing component. When it can be cleared, it is cleared directly by the obstacle clearing component. When it cannot be cleared, an alarm instruction is generated to send an alarm, thereby improving the stability and transportation efficiency of the transport robot.

Description

Technical Field

[0001] The present invention relates to the technical field of transport robots, in particular to an underground coal mine transport robot and a control system. Background Art

[0002] In the underground working environment of coal mines, transportation is a vital link. Traditional transportation methods mostly use rail cars or belt conveyors, but these methods have problems such as poor flexibility, complex operation and maintenance, and low safety. Especially in narrow tunnels, the turning radius of rail cars is large and cannot adapt to complex terrain. The installation and maintenance of belt conveyors require a lot of manpower and cannot respond in time when encountering emergencies such as landslides. In addition, the underground environment of coal mines is harsh, with unfavorable factors such as dust, high temperature, and high humidity, which puts extremely high demands on the reliability and durability of transportation equipment.

[0003] The Chinese patent with the publication number CN113580157A discloses a robot for transporting goods, comprising: a robot body, which includes a walking mechanism; a transport box, on which the first side wall of the two opposite side walls is placed on the bottom of the transport box, and the second side wall is rotatably connected to the bottom of the transport box and can be rotated toward the outside of the transport box until its top surface is flush with the top surface of the bottom of the transport box, and when the second side wall is parallel to the first side wall, the second side wall is detachably connected to at least one of the other two opposite side walls on the transport box; a storage plate, which is rotatably connected to the first side wall, and when the top surface of the second side wall is flush with the top surface of the bottom of the transport box, the first side wall and the storage plate can reciprocate in a direction perpendicular to the first side wall under the driving of power. The invention can facilitate loading and unloading by opening the second side wall and pushing the goods out through the storage plate, greatly improving the transportation efficiency.

[0004] With the development of automation technology, coal mine underground transportation systems based on robot technology have gradually become a research hotspot. In the existing technology, transportation robots are also applied to coal mine underground transportation operations. However, when goods fall in the coal mine or obstacles are caused by stones falling inside the mine, the existing transportation robots applied to coal mine underground are unable to perform intelligent analysis and processing of the obstacles during use. Generally, they will simply alarm or pass directly over the obstacles, which affects the stability of the transportation of the transportation robots and causes more goods to fall. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a coal mine underground transportation robot and a control system.

[0006] The present invention adopts the following technical scheme: a coal mine underground transport robot comprises a transport trolley and a carriage, the carriage is bolted to the upper surface of the transport trolley, a base is installed on one side of the carriage on the transport trolley, an obstacle removal component is arranged on the base, which is used to remove obstacles on the coal mine road, a bracket is also installed on the base, and a depth camera, an image analysis module and a controller are installed on the bracket;

[0007] The depth camera is used to obtain the depth image of obstacles on the transport path of the transport vehicle;

[0008] An image analysis module, which obtains the state coefficient of the obstacle based on the acquired obstacle depth image analysis, and generates an alarm instruction or an obstacle clearance instruction based on the state coefficient of the obstacle;

[0009] The controller is used to receive obstacle removal instructions to control the obstacle removal component to work and remove obstacles on the transportation path.

[0010] As a further description of the above technical solution: a method for obtaining the state coefficient of an obstacle based on the obtained obstacle depth image analysis includes:

[0011] ;

[0012] In the formula, is the state coefficient of the obstacle, is the obstacle height, is the obstacle volume, and is the weight coefficient.

[0013] As a further description of the above technical solution: the method for obtaining the obstacle height and obstacle volume includes:

[0014] Perform grayscale processing on the collected obstacle depth image, obtain the grayscale value of each pixel block in the image data, and mark it as a real-time grayscale value, and obtain the surface image data of the normal transportation path, obtain the grayscale value of each pixel block in the standard image data, and mark it as a standard grayscale value, preset a grayscale difference threshold, subtract the standard grayscale value of the same position in the standard image data from the real-time grayscale value of each block in the image data, and obtain the absolute value of the grayscale difference. When the absolute value of the grayscale difference is greater than or equal to the preset grayscale difference threshold, mark the pixel block as an obstacle, and when the absolute value of the grayscale difference is less than the preset grayscale difference threshold, do not mark the pixel block as an obstacle;

[0015] Extracting pixel blocks with obstacle marks, and mapping the positions of the pixel blocks with obstacle marks in the obstacle depth image to the blank background layer to obtain an extracted pattern composed of pixel blocks with obstacle marks;

[0016] Through the internal parameters of the depth camera, the depth value of each pixel is converted into a point in three-dimensional space to generate point cloud data;

[0017] Find the maximum value MAX(Z) and the minimum value Min(Z) on the Z axis in the point cloud data, and calculate the height H of the obstacle, H=MAX(Z)-Min(Z);

[0018] The volume of the obstacle is calculated by constructing the convex hull of the point cloud data, segmenting the convex hull using the triangulation method, and then calculating the volume of each triangle and summing them up to obtain the volume of the obstacle.

[0019] As a further description of the above technical solution: the method for generating an alarm stop instruction or an obstacle clearance instruction based on the state coefficient of the obstacle includes:

[0020] Preset state coefficient threshold ZWm;

[0021] when When ≤ZWm, a clearance instruction is generated;

[0022] when >ZWm, an alarm command is generated.

[0023] As a further description of the above technical solution: the obstacle removal component includes a strip guide rail, the strip guide rail is bolted to the front surface of the base, a sliding seat is slidably arranged in the strip guide rail, two positioning rods are plugged into the sliding seat, the bottom ends of the two positioning rods pass through the sliding seat and extend to the bottom and are welded with an obstacle removal scraper, a fixed seat is welded between the top ends of the two positioning rods, and an electric telescopic rod is installed between the fixed seat and the sliding seat, and the obstacle removal scraper is driven to adjust its height by controlling the extension and retraction of the electric telescopic rod, a lead screw is rotatably connected in the strip guide rail, the lead screw is threadedly connected to the sliding seat, and a lead screw motor that drives the lead screw to rotate is bolted on the outer wall of one end of the strip guide rail along the length direction;

[0024] The obstacle removal assembly also includes a bar-shaped support rail, which is fixedly connected to the transport trolley through a connecting seat. A support seat is slidably connected to the bar-shaped support rail, and a support rod is welded and fixed between the support seat and the sliding seat.

[0025] A coal mine underground transportation robot control system comprises the coal mine underground transportation robot and a control unit, wherein the control unit comprises:

[0026] The road segment division module obtains the transport path type composition of the transport vehicle, and divides the transport path into a transport segment according to different path types, thereby dividing the transport path into i transport segments, i>1, wherein the transport path types include the main transport lane, auxiliary transport lane, cut lane, ramp and inclined shaft segment;

[0027] A road section analysis module collects the road condition data of each transport section, wherein the road condition data of the transport section includes the IRI value, slope value, and curvature value of the transport section;

[0028] The transport analysis module collects the transport data of the carriage, the transport data including the stacking height and the transport weight, and generates a moving speed coefficient according to the road condition data of each transport section and the transport data of the carriage;

[0029] The moving speed prediction module converts the type of transport path and the moving speed coefficient into a set of feature vectors, and inputs the feature vectors into a pre-built predicted optimal moving speed machine learning model to obtain the predicted optimal moving speed, and then controls the moving speed of the transport vehicle based on the predicted optimal moving speed.

[0030] As a further description of the above technical solution: the method for generating a moving speed coefficient according to the road condition data of each transport section and the carriage carrying data includes:

[0031] ;

[0032] In the formula, Movement speed coefficient, is the IRI value, is the slope value, is the curvature value, is the stacking height, For carrying weight, , , , and is the weight coefficient, , , , and Both are greater than 0.

[0033] As a further description of the above technical solution: the training method of the machine learning model for predicting the optimal moving speed includes:

[0034] Collecting historical training data required for predicting the optimal moving speed in advance, wherein the historical training data is collected in an experimental environment, wherein the tester controls the physical factors that affect the increase or decrease of the moving speed of the transport trolley to obtain the optimal moving speed of the transport trolley, wherein the optimal moving speed is the fastest moving speed while ensuring the stable movement of the transport trolley;

[0035] The historical training data includes movement characteristic data and an optimal movement speed, wherein the movement characteristic data includes a type of transportation path and a movement speed coefficient;

[0036] Convert the collected historical training data into a corresponding set of feature vectors;

[0037] Each group of feature vectors is used as the input of the machine learning model, and the machine learning model takes the optimal moving speed corresponding to each group of mobile feature data as the output, and the optimal moving speed actually corresponding to each group of mobile feature data is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; training is stopped when the loss function value of the machine learning model is less than or equal to the preset target loss value.

[0038] As a further description of the above technical solution: the method for obtaining the curvature of the transport section includes:

[0039] Manually draw road paths using GIS tools;

[0040] Use the curvature analysis tool in the GIS software to calculate the curvature of each node of the road;

[0041] Calculate and obtain the average value of the curvature of all nodes to characterize the curvature of the transport section;

[0042] The method for obtaining the slope value of the transport section is to use a theodolite to measure the inclination angle and horizontal distance between two points on the road, and use a slope formula to calculate the slope.

[0043] As a further description of the above technical solution: the method for obtaining the IRI value of the transport section includes installing an IRI instrument on a test vehicle, the vehicle travels along the road at a preset speed, the IRI instrument records the data of the road surface profile, and then calculates the IRI value of the transport section through data processing software.

[0044] Beneficial effects:

[0045] The coal mine underground transport robot and control system provided by the present invention are equipped with an obstacle clearing component and a depth camera, which can collect images of obstacles on the transport path, obtain the state coefficient of the obstacle based on image analysis, and then judge whether the obstacle can be cleared by the obstacle clearing component. When the obstacle can be cleared, it is directly cleared by the obstacle clearing component. When it cannot be cleared, an alarm instruction is generated to send out an alarm, thereby improving the transportation stability and transportation efficiency of the transport robot, and overcoming the problem that the transport robot in the prior art cannot perform intelligent analysis and processing on obstacles, and generally only simply alarms or directly passes over the obstacle, affecting the transportation stability of the transport robot.

[0046] Furthermore, the transport path type composition of the transport vehicle is obtained, and then the transport path is independently divided into a transport section according to the different path types, and then an optimal moving speed is generated based on the path type, road condition status data and the carrying data of the transport vehicle of each transport section. That is, the transport vehicle can move at the optimal moving speed on different paths, and reasonable speed adjustment is achieved for different transport sections, overcoming the problem in the prior art that the transport vehicle can only travel on the entire moving path at the safe speed of the worst section, thereby improving the transport efficiency of the transport vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The present invention will be further explained below in conjunction with the accompanying drawings and embodiments:

[0048] Figure 1 The structure of the underground coal mine transport robot provided by the present invention is shown in FIG. Figure 1 ;

[0049] Figure 2 The structure of the underground coal mine transport robot provided by the present invention is shown in FIG. Figure 2 ;

[0050] Figure 3 A schematic diagram of the structure of the obstacle removal assembly provided by the present invention Figure 1 ;

[0051] Figure 4 A schematic diagram of the structure of the obstacle removal assembly provided by the present invention Figure 2 ;

[0052] Figure 5 This is a module connection diagram of the coal mine underground transportation robot control system provided by the present invention.

[0053] Figure numerals: 1. transport trolley; 2. carriage; 3. base; 4. bracket; 5. depth camera; 6. obstacle removal component; 61. strip guide rail; 62. sliding seat; 63. positioning rod; 64. obstacle removal scraper; 65. fixed seat; 66. electric telescopic rod; 67. screw; 68. screw motor; 69. connecting seat; 610. strip support rail; 611. support seat; 612. support rod. DETAILED DESCRIPTION

[0054] In order to make the technical means, creative features, objectives and effects of the present invention easy to understand, the present invention is further described below with reference to specific diagrams. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0055] Example 1

[0056] See also Figure 1-Figure 2The embodiment of the present invention provides a technical solution: a coal mine underground transport robot, comprising a transport trolley 1 and a carriage 2, the carriage 2 is bolted to the upper surface of the transport trolley 1, a base 3 is installed on one side of the carriage 2 on the transport trolley 1, an obstacle removal component 6 is arranged on the base 3, which is used to remove obstacles on the coal mine road, a bracket 4 is also installed on the base 3, and a depth camera 5, an image analysis module and a controller are installed on the bracket 4;

[0057] The depth camera 5 is used to obtain the depth image of obstacles on the transport path of the transport vehicle 1;

[0058] An image analysis module, which obtains the state coefficient of the obstacle based on the acquired obstacle depth image analysis, and generates an alarm instruction or an obstacle clearance instruction based on the state coefficient of the obstacle;

[0059] The controller is used to receive obstacle removal instructions to control the obstacle removal component 6 to work and remove obstacles on the transportation path.

[0060] The method for obtaining the state coefficient of the obstacle based on the obtained obstacle depth image analysis includes:

[0061] ;

[0062] In the formula, is the state coefficient of the obstacle, is the obstacle height, is the obstacle volume, and is the weight coefficient.

[0063] Specifically, obstacles on the transport path are cleared by the obstacle clearing component 6, and the clearing method is to push the obstacles to both sides of the transport path so as not to affect the movement of the transport vehicle 1. When the obstacle is high or large in size, the space on both sides of the transport path cannot bear the obstacle. At this time, the obstacle clearing process cannot be completed, and an alarm instruction needs to be generated for remote alarm.

[0064] It should be noted that the size of the weight coefficient is a specific value obtained by quantifying each data, which is convenient for subsequent comparison. The size of the weight coefficient depends on the number of comprehensive parameters and the preliminary setting of the corresponding weight coefficient for each group of comprehensive parameters by those skilled in the art;

[0065] The methods for obtaining obstacle height and obstacle volume include:

[0066] Perform grayscale processing on the collected obstacle depth image, obtain the grayscale value of each pixel block in the image data, and mark it as a real-time grayscale value, and obtain the surface image data of the normal transportation path, obtain the grayscale value of each pixel block in the standard image data, and mark it as a standard grayscale value, preset a grayscale difference threshold, subtract the standard grayscale value of the same position in the standard image data from the real-time grayscale value of each block in the image data, and obtain the absolute value of the grayscale difference. When the absolute value of the grayscale difference is greater than or equal to the preset grayscale difference threshold, mark the pixel block as an obstacle, and when the absolute value of the grayscale difference is less than the preset grayscale difference threshold, do not mark the pixel block as an obstacle;

[0067] Extracting pixel blocks with obstacle marks, and mapping the positions of the pixel blocks with obstacle marks in the obstacle depth image to the blank background layer to obtain an extracted pattern composed of pixel blocks with obstacle marks;

[0068] The depth value of each pixel is converted into a point in three-dimensional space through the internal parameters of the depth camera 5 to generate point cloud data;

[0069] Find the maximum value MAX(Z) and the minimum value Min(Z) in the Z-axis height direction in the point cloud data, and calculate the height H of the obstacle, H=MAX(Z)-Min(Z);

[0070] The volume of the obstacle is calculated by constructing the convex hull of the point cloud data, segmenting the convex hull using the triangulation method, and then calculating the volume of each triangle and summing them up to obtain the volume of the obstacle.

[0071] The method for generating an alarm stop instruction or an obstacle clearance instruction based on the state coefficient of the obstacle includes:

[0072] Preset state coefficient threshold ZWm;

[0073] when When ≤ZWm, a clearance instruction is generated;

[0074] when >ZWm, an alarm command is generated.

[0075] In this embodiment, the coal mine underground transport robot, when in use, when the transport trolley 1 moves on the transport path, can obtain the depth image of the obstacles on the transport path through the depth camera 5, obtain the state coefficient of the obstacles based on the obtained obstacle depth image analysis, and generate an alarm instruction or an obstacle clearance instruction based on the state coefficient of the obstacles. When the obstacle clearance instruction is generated, the obstacle clearance is automatically performed by the obstacle clearance component 6 provided on the transport trolley 1 to ensure the stability of the movement of the transport trolley 1. When the alarm instruction is generated, that is, the automatic obstacle clearance cannot be achieved by the obstacle clearance component 6 provided, the alarm instruction is generated and sent to the terminal PC to realize remote alarm. That is, the coal mine underground transport robot, which is equipped with the obstacle clearance component 6 and the depth camera 5, can collect the image of the obstacles on the transport path, and obtain the state coefficient of the obstacles based on the image analysis, and then judge whether the obstacles can be cleared by the obstacle clearance component 6. When the obstacles can be cleared, they are directly cleared by the obstacle clearance component 6. When the obstacles cannot be cleared, an alarm instruction is generated to issue an alarm.

[0076] Example 2

[0077] See also Figure 1-Figure 4 On the basis of the above-mentioned embodiments, the present embodiment further discloses the specific structure of the obstacle removal component 6. The obstacle removal component 6 comprises a strip guide rail 61, which is bolted to the front surface of the base 3. A sliding seat 62 is slidably arranged in the strip guide rail 61. Two positioning rods 63 are inserted on the sliding seat 62. The bottom ends of the two positioning rods 63 pass through the sliding seat 62 and extend to the bottom and are welded with an obstacle removal scraper 64. A fixed seat 65 is welded between the top ends of the two positioning rods 63, and an electric telescopic rod 66 is installed between the fixed seat 65 and the sliding seat 62. By controlling the extension and retraction of the electric telescopic rod 66, the obstacle removal scraper 64 is driven to adjust its height. A lead screw 67 is rotatably connected in the strip guide rail 61, and the lead screw 67 is threadedly connected to the sliding seat 62. A lead screw motor 68 for driving the lead screw 67 to rotate is bolted to the outer wall of one end of the strip guide rail 61 along the length direction.

[0078] Specifically, during use, when an obstacle that can be cleared is found in front of the transport trolley 1, the obstacle is cleared by the obstacle clearing component 6, and then the transport trolley 1 continues to transport. The specific obstacle clearing method of the obstacle clearing component 6 is to first control the electric telescopic rod 66 to retract, drive the obstacle clearing scraper 64 to move down to a position tangent to the road surface, and then control the screw motor 68 to work, drive the screw 67 to rotate, and through the rotation of the screw 67, drive the sliding seat 62 to move, and the sliding seat 62 drives the obstacle clearing scraper 64 to move through the positioning rod 63, push the obstacle to two points on the trolley's moving path, and then clean it regularly.

[0079] The obstacle removal component 6 also includes a bar support rail 610, which is fixedly connected to the transport trolley 1 through a connecting seat 69. A support seat 611 is slidably connected to the bar support rail 610, and a support rod 612 is welded and fixed between the support seat 611 and the sliding seat 62.

[0080] Specifically, the strip support rail 610 is provided to assist in supporting and limiting the sliding seat 62, thereby improving the stability of the sliding seat 62 in driving the obstacle removal scraper 64 to move and remove obstacles. When the sliding seat 62 moves, the support seat 611 slides on the strip support rail 610, so that during the movement of the sliding seat 62, the supporting rod 612 can be used to assist in supporting the sliding seat 62, thereby improving the strength of the sliding seat 62 and the stability of use.

[0081] Example 3

[0082] See also Figure 1-Figure 5 The embodiment of the present invention provides a technical solution: a coal mine underground transportation robot control system, including a coal mine underground transportation robot and a control unit, the control unit includes:

[0083] The road segment division module obtains the transport path type composition of the transport vehicle 1, and divides the transport path into a transport segment independently according to different path types, thereby dividing the transport path into i transport segments, i>1;

[0084] Among them, the types of transportation routes include sections of main transportation tunnels, auxiliary transportation tunnels, cutting tunnels, ramps and inclined shafts.

[0085] It should be noted that the main transport roadway is the most important transportation channel in the mine, usually used to transport coal, ore and other materials from the working face to the mine exit or hoist shaft. It is the backbone of the entire mine transportation network;

[0086] Auxiliary transport lanes are usually used for auxiliary transportation, mainly transporting equipment, materials and personnel. They connect the main transport lanes with various work areas or auxiliary facilities;

[0087] Cutting tunnels are short tunnels connecting the main transport tunnel or auxiliary transport tunnel with the coal mining face, mainly used for transporting coal and equipment;

[0088] A ramp is an inclined passage connecting tunnels at different heights, usually used to transport coal, ore and equipment;

[0089] Inclined shaft is used for inclined transportation and is one of the main channels connecting the ground and the underground;

[0090] The road section analysis module collects the road condition data of each transport section, and the road condition data of the transport section includes the IRI value, slope value and curvature value of the transport section;

[0091] Specifically, the method for obtaining the IRI value of the transport section includes installing an IRI instrument on a test vehicle, the vehicle driving along the road at a preset speed, the IRI instrument recording the data of the road surface profile, and then calculating the IRI value of the transport section through data processing software, wherein the IRI instrument senses the longitudinal profile data of the road, calculates the response of the vehicle suspension system, and then calculates the flatness of the road. It should be noted that the larger the IRI value, the more uneven the road is;

[0092] The method for obtaining the slope value of the transport section is to use a theodolite to measure the inclination angle and horizontal distance between two points on the road, and use the slope formula to calculate the slope.

[0093] Methods for obtaining the curvature of a transport section include:

[0094] Manually draw road paths using GIS tools;

[0095] Use the curvature analysis tool in the GIS software to calculate the curvature of each node of the road;

[0096] The average value of the curvature of all nodes is calculated to characterize the curvature of the transport section.

[0097] The transport analysis module collects the transport data of carriage 2, which includes the stacking height and the transport weight, and generates a moving speed coefficient according to the road condition data of each transport section and the transport data of carriage 2.

[0098] The moving speed prediction module converts the type of transport path and the moving speed coefficient into a set of feature vectors, and inputs the feature vectors into a pre-built predicted optimal moving speed machine learning model to obtain the predicted optimal moving speed, and then controls the moving speed of the transport cart 1 based on the predicted optimal moving speed.

[0099] The method for generating the moving speed coefficient according to the road condition data of each transport section and the carrying data of the carriage 2 includes:

[0100] ;

[0101] In the formula, Movement speed coefficient, is the IRI value, is the slope value, is the curvature value, is the stacking height, For carrying weight, , , , and is the weight coefficient, , , , and Both are greater than 0.

[0102] Specifically, the larger the IRI value, the more uneven the road is, and the lower the moving speed is, and vice versa. The larger the slope value, the lower the moving speed is, and vice versa. The larger the curvature value, the lower the moving speed is, and vice versa. The larger the loading height, the lower the moving speed is, and vice versa. The larger the load weight, the lower the moving speed is. In summary, the larger the moving speed coefficient, the lower the moving speed of the transport vehicle 1 is, and vice versa.

[0103] It should be noted that the size of the weight coefficient is a specific value obtained by quantifying each data to facilitate subsequent comparison. The size of the weight coefficient depends on the number of comprehensive parameters and the preliminary setting of the corresponding weight coefficient for each set of comprehensive parameters by technical personnel in this field.

[0104] The training method of the machine learning model for predicting the optimal moving speed includes:

[0105] Collect historical training data required for predicting the optimal moving speed in advance. The historical training data is collected in an experimental environment. The experimental environment is a tester controlling the physical factors that affect the increase or decrease of the moving speed of the transport trolley 1 to obtain the optimal moving speed of the transport trolley 1. The optimal moving speed is the fastest moving speed while ensuring the stable movement of the transport trolley 1.

[0106] The historical training data includes movement characteristic data and optimal movement speed, and the movement characteristic data includes the type of transportation path and the movement speed coefficient;

[0107] Convert the collected historical training data into a corresponding set of feature vectors;

[0108] Each set of feature vectors is used as the input of the machine learning model. The machine learning model takes the optimal moving speed corresponding to each set of mobile feature data as the output, the optimal moving speed actually corresponding to each set of mobile feature data is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; training is stopped when the loss function value of the machine learning model is less than or equal to the preset target loss value.

[0109] It should be noted that the machine learning model can be one of the models such as support vector machine regression, random forest regression or neural network regression.

[0110] The loss function value of the machine learning model is the mean square error.

[0111] Mean square error is one of the commonly used loss functions. Minimization is used to train the model so that the machine learning model can better fit the data, thereby improving the performance and accuracy of the model.

[0112] In the loss function is the loss function value of the machine learning model, is the feature vector group number; is the number of eigenvector groups; For the The optimal moving speed corresponding to the group eigenvector, For the The optimal moving speed corresponding to the group feature vector in real time.

[0113] Other model parameters of the machine learning model, target loss value, optimization algorithm, training set test set validation set ratio, and loss function optimization are all achieved through actual engineering implementation and continuous experimental tuning.

[0114] In this embodiment, the transport path type composition of the transport vehicle 1 is obtained, and then the transport path is independently divided into a transport section according to different path types, and then an optimal moving speed is generated based on the path type, road condition status data and the carrying data of the transport vehicle 1 of each transport section. That is, the transport vehicle 1 can move at the optimal moving speed on different paths, and reasonable speed adjustment is achieved for different transport sections, overcoming the problem in the prior art that the transport vehicle 1 can only travel on the entire moving path at the safe speed of the worst section, thereby improving the transport efficiency of the transport vehicle 1.

[0115] The basic principles, main features and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and the description in the specification are only to illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A coal mine underground transport robot, characterized in that: The invention comprises a transport trolley (1), an obstacle removal component (6) and a carriage (2), wherein the carriage (2) is bolted to the upper surface of the transport trolley (1), a base (3) is installed on one side of the carriage (2) on the transport trolley (1), the obstacle removal component (6) is arranged on the base (3) and is used to remove obstacles on the road in the coal mine, and a bracket (4) is also installed on the base (3), and a depth camera (5), an image analysis module and a controller are installed on the bracket (4); The depth camera (5) is used to obtain a depth image of obstacles on the transport path of the transport vehicle (1); An image analysis module, which obtains the state coefficient of the obstacle based on the acquired obstacle depth image analysis, and generates an alarm instruction or an obstacle clearance instruction based on the state coefficient of the obstacle; A controller, used for receiving an obstacle removal instruction to control the obstacle removal component (6) to operate and remove obstacles on the transport path; The method for obtaining the state coefficient of the obstacle based on the obtained obstacle depth image analysis includes: ； In the formula, is the state coefficient of the obstacle, is the obstacle height, is the obstacle volume, and is the weight coefficient; The method for obtaining the obstacle height and obstacle volume includes: Perform grayscale processing on the collected obstacle depth image, obtain the grayscale value of each pixel block in the image data, and mark it as a real-time grayscale value, and obtain the surface image data of the normal transportation path, obtain the grayscale value of each pixel block in the standard image data, and mark it as a standard grayscale value, preset a grayscale difference threshold, subtract the standard grayscale value of the same position in the standard image data from the real-time grayscale value of each block in the image data, and obtain the absolute value of the grayscale difference. When the absolute value of the grayscale difference is greater than or equal to the preset grayscale difference threshold, mark the pixel block as an obstacle, and when the absolute value of the grayscale difference is less than the preset grayscale difference threshold, do not mark the pixel block as an obstacle; Extracting pixel blocks with obstacle marks, and mapping the positions of the pixel blocks with obstacle marks in the obstacle depth image to the blank background layer to obtain an extracted pattern composed of pixel blocks with obstacle marks; The depth value of each pixel is converted into a point in three-dimensional space through the internal parameters of the depth camera (5) to generate point cloud data; Find the maximum value MAX(Z) and the minimum value Min(Z) in the Z-axis height direction in the point cloud data, and calculate the height H of the obstacle, H=MAX(Z)-Min(Z); The volume of the obstacle is calculated by constructing the convex hull of the point cloud data, segmenting the convex hull using the triangulation method, and then calculating the volume of each triangle and summing them up to obtain the volume of the obstacle; The method for generating an alarm stop instruction or an obstacle clearance instruction based on the state coefficient of the obstacle includes: Preset state coefficient threshold ZWm; when When ≤ZWm, a clearance instruction is generated; when >ZWm, an alarm command is generated.

2. The coal mine underground transportation robot according to claim 1, characterized in that: The obstacle removal component (6) comprises a strip guide rail (61), the strip guide rail (61) being bolted to the front surface of the base (3), a sliding seat (62) being slidably arranged in the strip guide rail (61), two positioning rods (63) being inserted into the sliding seat (62), the bottom ends of the two positioning rods (63) passing through the sliding seat (62) extending downward and being welded with an obstacle removal scraper (64), a fixed seat (65) being welded between the top ends of the two positioning rods (63), and an electric telescopic rod (66) being installed between the fixed seat (65) and the sliding seat (62), the height of the obstacle removal scraper (64) being driven to be adjusted by controlling the electric telescopic rod (66) to be extended and retracted, a lead screw (67) being rotatably connected in the strip guide rail (61), the lead screw (67) being threadedly connected to the sliding seat (62), and a lead screw motor (68) for driving the lead screw (67) to rotate is bolted to the outer wall of one end of the strip guide rail (61) along the length direction; The obstacle removal assembly (6) further comprises a strip-shaped support rail (610), wherein the strip-shaped support rail (610) is fixedly connected to the transport trolley (1) via a connecting seat (69), a support seat (611) is slidably connected to the strip-shaped support rail (610), and a support rod (612) is welded and fixed between the support seat (611) and the sliding seat (62).

3. A coal mine underground transportation robot control system, comprising a coal mine underground transportation robot and a control unit as claimed in any one of claims 1 to 2, characterized in that: The control unit comprises: A road segment division module, which obtains the transport path type composition of the transport vehicle (1), and divides the transport path into a transport section independently according to different path types, thereby dividing the transport path into i transport sections, i>1, wherein the transport path types include sections of a main transport lane, an auxiliary transport lane, a cut lane, a ramp, and an inclined shaft; A road section analysis module collects the road condition data of each transport section, wherein the road condition data of the transport section includes the IRI value, the slope value and the curvature value of the transport section; A transport analysis module collects transport data of the carriage (2), the transport data including a stacking height and a transport weight, and generates a moving speed coefficient based on the road condition data of each transport section and the transport data of the carriage (2); The moving speed prediction module converts the type of transport path and the moving speed coefficient into a set of feature vectors, and inputs the feature vectors into a pre-built machine learning model for predicting the optimal moving speed to obtain the predicted optimal moving speed, and then controls the moving speed of the transport vehicle (1) based on the predicted optimal moving speed.

4. The coal mine underground transportation robot control system according to claim 3, characterized in that: The training method of the machine learning model for predicting the optimal moving speed includes: Collecting in advance historical training data required for predicting the optimal moving speed, the historical training data being collected in an experimental environment, wherein the experimental environment is a tester controlling physical factors that affect the increase or decrease of the moving speed of the transport trolley (1) to obtain the optimal moving speed of the transport trolley (1), wherein the optimal moving speed is the fastest moving speed while ensuring the stable movement of the transport trolley (1); The historical training data includes movement characteristic data and an optimal movement speed, wherein the movement characteristic data includes a type of transportation path and a movement speed coefficient; Convert the collected historical training data into a corresponding set of feature vectors; Each group of feature vectors is used as the input of the machine learning model, and the machine learning model takes the optimal moving speed corresponding to each group of mobile feature data as the output, and the optimal moving speed actually corresponding to each group of mobile feature data is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; training is stopped when the loss function value of the machine learning model is less than or equal to the preset target loss value.

5. The coal mine underground transportation robot control system according to claim 3, characterized in that: The method for obtaining the curvature of the transport section includes: Manually draw road paths using GIS tools; Use the curvature analysis tool in the GIS software to calculate the curvature of each node of the road; Calculate and obtain the average value of the curvature of all nodes to characterize the curvature of the transport section; The method for obtaining the slope value of the transport section is to use a theodolite to measure the inclination angle and horizontal distance between two points on the road, and use a slope formula to calculate the slope.

6. The coal mine underground transportation robot control system according to claim 3, characterized in that: The method for obtaining the IRI value of the transport section includes installing an IRI instrument on a test vehicle, the vehicle driving along the road at a preset speed, the IRI instrument recording data of the road surface profile, and then calculating the IRI value of the transport section through data processing software.

Images