Automatic rod changing mechanism of coal mine drilling and anchoring robot and precision compensation method of automatic rod changing mechanism

Abstract

The invention provides an automatic rod changing mechanism of a coal mine drilling and anchoring robot and a precision compensation method of the automatic rod changing mechanism, and relates to the technical field of rod changing of the coal mine drilling and anchoring robotic.The automatic rod changing mechanism comprises a mounting plate, a six-axis mechanical arm, a servo pneumatic clamping jaw and a microcontroller, an anchor rod storage barrel is mounted above the mounting plate through a supporting frame, and an anchor rod outlet is formed in one end of the anchor rod storage barrel; a plurality of groups of discs are uniformly arranged on the inner side of the anchor rod storage barrel, the plurality of groups of discs are connected with the anchor rod storage barrel through a driving assembly, a first opening is formed in the upper end of the anchor rod storage barrel, an L-shaped assembly plate is mounted on the mounting plate, an anchor rod output assembly is mounted on the L-shaped assembly plate, and anchor rod bayonets are formed in the discs; by integrating the precision compensation learning model and the deep reinforcement learning technology, the problem that the output precision of the anchor rod in an existing automatic rod replacing system is insufficient is solved. Through accurate control and real-time error compensation, smooth and accurate output of the anchor rod is ensured, and the automation level in coal mine operation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of rod changing for coal mine drilling and anchoring robots, and in particular, to an automatic rod changing mechanism for coal mine drilling and anchoring robots and a precision compensation method therefor. Background Art

[0002] With the rapid development of science and technology, the intelligentization of coal mining has become an important development direction. In recent years, the pace of automation and intelligentization in fully mechanized mining faces has been accelerating continuously, posing severe challenges to the production capacity and advancing speed of fully mechanized heading faces, which requires the speed of excavation support to also increase accordingly. Currently, most coal mines in China have partially achieved mechanization in excavation support operations. Roadways are excavated using roadheaders, and support is carried out using drilling and anchoring machinery or integrated drilling and anchoring machines. Basically, it is in a state of personnel following the machine for operation. However, the drilling and anchoring speed seriously lags behind the excavation speed, the working environment is poor, the labor intensity is high, the excavation efficiency is low, and it cannot meet the progress requirements of fully mechanized mining, directly affecting the safe, high-yield, and efficient coal mining.

[0003] As an integrated automated device, coal mine drilling and anchoring robots are widely used in coal mine mining, tunnel engineering, and reinforcement operations of underground structures. Its main function is to automatically complete drilling, bolt installation, and support work, reducing manual operation and improving work efficiency and safety. Coal mine drilling and anchoring robots usually consist of core components such as robotic arms, servo motors, control systems, and sensors, and can complete high-precision drilling and bolt installation tasks.

[0004] In the application of coal mine drilling and anchoring robots, the automatic rod changing system is an important part of improving work efficiency, reducing manual intervention, and enhancing operation safety. Coal mine drilling and anchoring robots can automatically complete drilling and bolt installation tasks, greatly improving the efficiency and safety of coal mine support operations. However, in existing automatic rod changing technologies, the accuracy of bolt output usually depends on the fixed structure of mechanical components and simple sensor feedback. This structural design cannot effectively compensate for the motion errors of mechanical components, resulting in the inability of bolts to accurately output from the bolt storage cylinder to the target position. Since the system cannot adjust the output process according to real-time data, situations such as bolts getting stuck, misaligned, or unable to be smoothly transported to the grasping area of the robotic arm often occur. At the same time, the lack of a real-time precision compensation mechanism during the bolt output process leads to inaccurate bolt output positions, thus affecting the accuracy of subsequent operations. Therefore, the present invention proposes an automatic rod changing mechanism for coal mine drilling and anchoring robots and a precision compensation method therefor to solve the problems existing in the prior art. Summary of the Invention

[0005] In view of the above problems, the object of the present invention is to propose an automatic rod-changing mechanism for a coal mine drilling and anchoring robot and its accuracy compensation method. The automatic rod-changing mechanism for a coal mine drilling and anchoring robot and its accuracy compensation method have the advantage of improving the output accuracy of anchor rods during the automatic rod-changing process and can solve the problems existing in the prior art.

[0006] To achieve the object of the present invention, the present invention is realized through the following technical solutions: An automatic rod-changing mechanism for a coal mine drilling and anchoring robot includes a mounting plate, a six-axis robotic arm, a servo pneumatic gripper, and a microcontroller. A servo linear module is installed on the mounting plate, and the six-axis robotic arm is connected to the servo linear module through a fixing plate. A servo pneumatic gripper is installed on the six-axis robotic arm. Above the mounting plate, an anchor rod storage cylinder is installed through a support frame, and one end of the anchor rod storage cylinder is provided with an anchor rod outlet. A disc is arranged inside the anchor rod storage cylinder, and several groups of discs are evenly arranged. The several groups of discs are connected to the anchor rod storage cylinder through a driving component. The upper end of the anchor rod storage cylinder is provided with a first opening. An L-shaped assembly plate is installed on the mounting plate, and an anchor rod output component is installed on the L-shaped assembly plate, and the position of the anchor rod output component corresponds to the first opening. Anchor rod clamping openings are arranged on the disc, and several groups of anchor rod clamping openings are evenly arranged.

[0007] Further improvement lies in that: The anchor rod output component includes several groups of servo cylinders. The output ends of the servo cylinders pass through the L-shaped assembly plate and are installed with a fixing frame. A rubber roller is installed on the fixing frame through a bearing, and the rubber roller is driven by a hub motor.

[0008] Further improvement lies in that: One end of the anchor rod clamping opening is arc-shaped, and the other end of the anchor rod clamping opening communicates with the outside. Universal balls are installed inside the anchor rod clamping opening, and several groups of universal balls are evenly arranged.

[0009] Further improvement lies in that: The driving component includes a main shaft. Both ends of the main shaft are connected to the anchor rod storage cylinder through bearings, and the main shaft is driven by a servo motor. An extension block is installed on the main shaft. A card slot is arranged inside the disc. The number and position of the card slot and the extension block are the same, and the card slot is adapted to the extension block.

[0010] Further improvement lies in that: An extension rod is installed on the mounting plate, and an arc-shaped positioning block is installed at the upper end of the extension rod. The lower vertex of the arc-shaped positioning block and the lower vertex of the anchor rod outlet are on the same horizontal line.

[0011] Further improvement lies in that: Protective rubber pads are installed at both ends inside the anchor rod storage cylinder.

[0012] An accuracy compensation method for an automatic rod-changing mechanism of a coal mine drilling and anchoring robot includes the following steps:

[0013] Step 1: Data collection

[0014] Collect the rotational angle data of the disc, the contact pressure data between the rubber roller and the anchor rod, the speed data of the servo linear module and the servo cylinder, and the operation position data of the six-axis robotic arm by means of arranging sensors;

[0015] Step 2: Construct the accuracy compensation learning model

[0016] Construct an accuracy compensation learning model based on the random forest regression model and the deep reinforcement learning method, and use the collected sensor data and historical operation data for training, so that the accuracy compensation learning model compensates the detected error according to the current sensor data;

[0017] Step 3: Deployment and use of the accuracy compensation learning model

[0018] Deploy the accuracy compensation learning model into the microcontroller, process the collected real-time data in real time through the microcontroller, output the prediction result, and then dynamically adjust the accuracy compensation according to the result;

[0019] Step 4: Feedback and optimization of accuracy compensation

[0020] Collect feedback on the accuracy compensation effect through the intervention of an external operator, and optimize the accuracy compensation learning model.

[0021] The further improvement lies in: in the fourth step, the specific steps of optimization are:

[0022] S1. Collect the feedback data of the external operator and label it;

[0023] S2. Integrate the labeled data into the training set of the model as calibration data;

[0024] S3. Through the deep reinforcement learning method, the accuracy compensation learning model uses the calibration data for corrective training to complete the optimization.

[0025] The beneficial effects of the present invention are:

[0026] (1) By integrating key components such as servo cylinders, rubber rollers, and disc rotation control, and combining with the accuracy compensation learning model, this device realizes precise control of the output accuracy of the anchor rod. Especially when the anchor rod is output, it can ensure that each anchor rod accurately passes through the first opening and smoothly enters the grasping area of the robotic arm through real-time sensor feedback and accuracy compensation mechanism, avoiding problems such as anchor rod output failure or jamming.

[0027] (2) This device adopts deep reinforcement learning and a random forest regression model, which can compensate for errors in the anchor rod output process in real time. By collecting data such as the rotation angle of the disc and the pressure of the rubber roller in real time, it can dynamically adjust the action parameters of components such as the servo motor and the cylinder, compensate for the influence of mechanical errors and environmental changes, and ensure the accuracy and stability of the anchor rod output process. At the same time, it can automatically optimize the control strategy according to the feedback data after each operation, thereby improving the accuracy and operational stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a front view structural schematic diagram of the present invention.

[0029] Figure 2 is a side view schematic diagram of the present invention.

[0030] Figure 3 is a side view structural schematic diagram of the anchor rod storage cylinder of the present invention.

[0031] Figure 4 is a front view structural schematic diagram after the anchor rod is removed of the present invention.

[0032] Figure 5 is a top view schematic diagram of the linear module distribution of the present invention.

[0033] Wherein: 1. mounting plate; 2. six-axis robotic arm; 3. servo pneumatic gripper; 4. servo linear module; 5. anchor rod storage cylinder; 6. anchor rod outlet; 7. disc; 8. first opening; 9. L-shaped assembly plate; 10. anchor rod bayonet; 11. servo cylinder; 12. fixing bracket; 13. rubber roller; 14. universal ball; 15. main shaft; 16. servo motor; 17. extension block; 18. card slot; 19. extension rod; 20. arc-shaped positioning block; 21. protective rubber pad. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] In order to deepen the understanding of the present invention, the following will further elaborate on the present invention in combination with embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0035] Traditional automatic rod-changing systems lack a real-time accuracy compensation mechanism during the anchor rod output process. Driving devices such as cylinders and servo motors are prone to performance deviations after long-term operation, and existing systems cannot make dynamic adjustments according to the feedback of sensors, resulting in inaccurate anchor rod output positions and affecting the accuracy of subsequent operations. Especially in complex working environments, external factors such as equipment wear and sensor drift often lead to a reduction in output accuracy.

[0036] Meanwhile, in the prior art, errors are likely to occur during the movement of mechanical components (such as discs, cylinders, grippers, etc.). These errors may accumulate over time, affecting the output accuracy of the anchor rod. Especially when multiple mechanical components work together, subtle deviations may be amplified, resulting in an increase in the overall output error, thus affecting the stability and efficiency of the entire rod-changing process.

[0037] Therefore, according to Figures 1 - 5 As shown, this embodiment proposes an automatic rod-changing mechanism for a coal mine drilling and anchoring robot. For this device, the main way to replace the anchor rod depends on the six-axis robotic arm 2 and the servo pneumatic gripper 3. A brief description of the steps is as follows:

[0038] First, the six-axis robotic arm 2 first locates to the drilling position of the coal mine drilling and anchoring robot, uses the servo pneumatic gripper 3 to grasp the old anchor rod in the current drill hole, then pulls it out and transfers it to the corresponding waste anchor rod storage area. Then, a new anchor rod is taken out from the anchor rod storage cylinder 5, and the six-axis robotic arm 2 installs it to the current drill hole position, thus completing an automatic rod-changing process. During the accuracy compensation process of this device, it mainly involves the positioning accuracy of the new anchor rod output, the contact accuracy of the new anchor rod transportation, and the grasping accuracy of the six-axis robotic arm 2.

[0039] Furthermore, this device includes a mounting plate 1, a six-axis robotic arm 2, a servo pneumatic gripper 3, and a microcontroller. The microcontroller corresponds to a microcomputer, which is responsible for electrically connecting to all the electronic components in this device and then controlling them.

[0040] A servo linear module 4 is installed on the mounting plate 1, and the six-axis robotic arm 2 is connected to the servo linear module 4 through a fixing plate. The servo pneumatic gripper 3 is installed on the six-axis robotic arm 2. The servo linear module 4 consists of a high-precision servo motor and a precision guide rail. Its function is to provide precise linear motion. Its sliding end is fixedly connected to the base of the six-axis robotic arm 2, which can meet the motion requirements of the six-axis robotic arm 2 and the servo pneumatic gripper 3, ensuring that the robot can complete the rod-changing operation in three-dimensional space. Specifically, the servo pneumatic gripper 3 is bolted to the end effector position of the six-axis robotic arm 2 to ensure that the gripper can accurately perform grasping and releasing operations along with the movement of the six-axis robotic arm 2. The mounting plate 1 is used to fix to the coal mine drilling and anchoring robot.

[0041] Above the mounting plate 1, a bolt storage cylinder 5 is installed through a support frame, and one end of the bolt storage cylinder 5 is provided with a bolt outlet 6 (the right end in this embodiment). In this embodiment, the bolt outlet 6 can only accommodate a group of bolts to enter and exit, ensuring the stable output of the bolts and the precise grasping of the six-axis robotic arm 2. Correspondingly, a disc 7 is provided inside the bolt storage cylinder 5, and several groups of discs 7 are evenly arranged. The function of the disc 7 is to separate the bolts in the bolt storage cylinder 5. In this embodiment, five groups of discs 7 are evenly distributed. At the same time, bolt clamping openings 10 are provided on the disc 7, and several groups of bolt clamping openings 10 are evenly arranged. The diameter of the disc 7 is smaller than the inner diameter of the bolt storage cylinder 5. Thus, the design of the disc 7 enables each group of bolts to be stably stored in the bolt clamping openings 10 of each disc 7, avoiding the instability of the bolts due to excessive space. Several groups of discs 7 are connected to the bolt storage cylinder 5 through a drive assembly. The drive assembly includes a main shaft 15. Both ends of the main shaft 15 are connected to the bolt storage cylinder 5 through bearings, and the main shaft 15 is driven by a servo motor 16. The servo motor 16 is located outside the bolt storage cylinder 5 (the end far from the bolt outlet 6). An extension block 17 is installed on the main shaft 15. The extension block 17 is integrally formed with the main shaft 15. A card slot 18 is provided inside the disc 7. The number and position of the card slot 18 and the extension block 17 are the same, and the card slot 18 is adapted to the extension block 17. In this embodiment, six card slots 18 are provided inside each group of discs 7, and then an extension block 17 is correspondingly provided in each card slot 18. Further, the extension block 17 is fixedly connected to the disc 7. By providing the extension block 17, the contact area between the main shaft 15 and the disc 7 is increased, thereby improving its stability and ensuring the position accuracy of each disc 7 after installation, thus greatly improving the reliability and accuracy of the rod-changing process.

[0042] The upper end of the bolt storage cylinder 5 is provided with a first opening 8. An L-shaped assembly plate 9 is installed on the mounting plate 1, and one end of the L-shaped assembly plate 9 extends directly above the bolt storage cylinder 5. A bolt output assembly is installed on the L-shaped assembly plate 9, and the position of the bolt output assembly corresponds to the first opening 8. The bolt output assembly includes a servo cylinder 11. A number of groups of servo cylinders 11 are evenly arranged. The output end of the servo cylinder 11 passes through the L-shaped assembly plate 9 and is installed with a fixing frame 12. A rubber roller 13 is installed on the fixing frame 12 through a bearing, and the rubber roller 13 is driven by a hub motor. The hub motor provides power through motor control to drive the rubber roller 13 to rotate, thereby generating friction to push the bolt to move along a specified path towards the bolt outlet 6. Further, during operation, the speed and torque of the motor precisely control the speed of the rubber roller to ensure that the bolt can be smoothly output at an appropriate speed. In this device, each time a group of bolts is conveyed. The bolts to be conveyed are rotated by the rotating disk 7 to directly below the first opening 8, and then the servo cylinder 11 drives the fixing frame 12 to move downward, so as to prompt the rubber roller 13 to contact the bolt (by adjusting the output force and stroke of the servo cylinder 11 cylinder, the position and contact pressure of the rubber roller 13 can be precisely controlled). In this device, one end of the bolt bayonet 10 is arc-shaped, and the other end of the bolt bayonet 10 communicates with the outside. A number of groups of universal balls 14 are evenly installed inside the bolt bayonet 10. The function of the universal balls 14 is to facilitate the movement of the bolt. Then the rubber roller 13 starts, and the friction generated by its contact drives the bolt to move out from the bolt outlet 6. At this time, the six-axis robotic arm 2 drives the servo pneumatic gripper 3 to enter the established grasping area to grasp the moved-out bolt. The servo pneumatic gripper 3 can adjust the clamping force through pneumatic control to ensure safety and stability during the grasping process.

[0043] Correspondingly, through the coordinated work of key components such as the bolt storage cylinder 5, the disk 7, the servo cylinder 11, and the rubber roller 13, this device realizes an accurate bolt output process. Each time the output bolt is moved to below the first opening 8 by the rotating disk 7 in a rotating state, pushed by the servo cylinder 11 and the rubber roller 13, and finally grasped by the six-axis robotic arm 2.

[0044] An extension rod 19 is installed on the mounting plate 1, and an arc-shaped positioning block 20 is installed at the upper end of the extension rod 19. The lower vertex of the arc-shaped positioning block 20 is on the same horizontal line as the lower vertex of the bolt outlet 6. When the bolt moves out a certain distance, it contacts the arc-shaped positioning block 20, playing a role in precise guidance.

[0045] Further, protective rubber pads 21 are installed at both ends inside the bolt storage cylinder 5 to prevent the bolts from directly colliding with both ends of the bolt storage cylinder 5.

[0046] An accuracy compensation method for the automatic rod-changing mechanism of a coal mine drilling and anchoring robot, comprising the following steps:

[0047] Step 1. Data acquisition (Data of the entire rod-changing mechanism is acquired through arranged sensors to obtain key parameters and real-time feedback, so as to accurately perform accuracy compensation)

[0048] By means of arranging sensors, collect the rotation angle data of the disc 7, the contact pressure data between the rubber roller 13 and the anchor rod, the speed data of the servo linear module 4 and the servo cylinder 11, and the operation position data of the six-axis robotic arm 2.

[0049] Specifically, install a rotary encoder on the disc 7 to monitor the rotation angle of the disc in real time. The rotary encoder can provide high-precision angle data, ensure that there is no error in the rotation process of the disc, so as to ensure that the anchor rod can accurately reach the predetermined position during each rod change. This data is used to judge the current rotation state of the disc, ensure that the rotation angle of the disc 7 has been adjusted in place when each anchor rod is output, and make preparations for subsequent operations.

[0050] Install a pressure sensor (such as a strain-type pressure sensor or a piezoelectric sensor) in the area where the rubber roller 13 contacts the anchor rod to monitor the contact pressure between the two in real time. The pressure sensor can detect the change of the contact force and adjust the pressure of the roller through the feedback data to avoid affecting the conveyance of the anchor rod due to too large or too small pressure. This data is used to understand the friction force between the roller and the anchor rod, ensure that the pressure is always controlled within the optimal range during the output process of the anchor rod, so as to avoid damage or smooth sliding of the anchor rod.

[0051] Install speed sensors or encoders on the servo linear module 4 and the servo cylinder 11 to monitor the speed and motion state of these two driving devices in real time. These sensors can provide linear displacement data and speed data, help the control device accurately understand the motion speed and position of these key components. By monitoring the speed data of the servo linear module 4 and the servo cylinder 11, the motion trajectory and control strategy can be dynamically adjusted to ensure that the roller 13 or other execution components move along the predetermined path and avoid errors or delays.

[0052] Install a vision sensor on the six-axis robotic arm 2 to obtain the operation position data of the robotic arm in real time. The position sensor can provide the accurate position and posture of the end of the robotic arm, ensure the positioning accuracy when grasping the anchor rod, and is used to accurately control the grasping action of the robotic arm, ensure that the robotic arm can accurately dock at each key step during the conveyance process of the anchor rod, and avoid position errors when performing the grasping task.

[0053] Step 2. Build an accuracy compensation learning model

[0054] Construct an accuracy compensation learning model based on a random forest regression model and a deep reinforcement learning method, and train it using the collected sensor data and historical operation data, so that the accuracy compensation learning model compensates for the detected error according to the current sensor data.

[0055] Among them, random forest regression is an ensemble learning method that makes predictions by generating multiple decision trees and outputs the prediction results through the average of the trees. In the accuracy compensation model, the application of the random forest regression model can help predict the possible errors of each component (such as disks, cylinders, robotic arms, etc.) under given conditions. Correspondingly, using the sensor data (such as disk rotation angle, pressure, speed, robotic arm position, etc.) collected in Step 1 as input features, through training, the random forest regression model can predict the possible errors of the device according to the input data (current sensor feedback). For example, if the position of the robotic arm deviates from the predetermined trajectory, the model will predict the specific value of the error and provide compensation information. Therefore, during the training process, historical operation data and known error values are used to train the random forest model. The historical data should contain error information under various operating conditions (for example, when the environmental temperature changes or the sensor fails, the possible errors of the device). The trained model can predict the accuracy error of the device according to the real-time input data and provide corresponding compensation suggestions.

[0056] Deep reinforcement learning is a technology that combines deep learning and reinforcement learning. It can optimize strategies through interaction with the environment. Then, during the training process, by simulating the device operation environment, the agent learns the optimal strategy in each interaction with the environment and optimizes the compensation decision through the reward function. The agent will gradually learn how to reduce errors and improve accuracy through multiple trials and errors. Correspondingly, by combining the random forest regression model and deep reinforcement learning, an efficient accuracy compensation learning model can be constructed. The random forest regression model is used to predict errors, and the deep reinforcement learning method optimizes the compensation strategy through adaptive adjustment. Then, it can perform precise error compensation according to real-time sensor data, improving the accuracy and stability during the rod-changing process. At the same time, it can self-learn and optimize in a changing working environment, thus providing long-term and continuous accuracy guarantee.

[0057] Step 3: Deployment and Use of the Accuracy Compensation Learning Model

[0058] Deploy the accuracy compensation learning model into the microcontroller. The microcontroller processes the collected real-time data in real time and outputs the prediction results, and then dynamically adjusts the accuracy compensation according to the results. The microcontroller is an ARM Cortex series processor, which has strong computing power and high-speed data processing capabilities. Therefore, embed the accuracy compensation learning model into the microcontroller to process sensor data in real time and generate compensation outputs, and finally perform control.

[0059] Correspondingly, the microcontroller receives the feedback data from each sensor in real time, performs necessary preprocessing (denoising, normalization, and smoothing) on this data, and then inputs it into the model. According to the prediction results of the model, the microcontroller will dynamically adjust the control parameters of this device. Specifically, the control device will perform precision compensation adjustment through execution components such as servo motor 16, servo cylinder 11, and six-axis robotic arm 2 to ensure that parameters such as the movement, pressure, and speed of this device during the automatic rod-changing process always remain in the optimal state. If the prediction results show that the rotation angle deviation of the disc 7 is large, the microcontroller will adjust the speed or torque of the servo motor 16 to ensure that the disc 7 rotates accurately to the predetermined position; if the prediction results show that the pressure between the rubber roller 13 and the anchor rod is too large or too small, the microcontroller will adjust the output of the servo cylinder 11 to ensure that the pressure is always controlled within the ideal range, thereby avoiding damage to the anchor rod or smooth transportation.

[0060] Step Four: Feedback and Optimization of Precision Compensation

[0061] Through the intervention of an external operator, collect feedback on the precision compensation effect and optimize the precision compensation learning model. The specific steps for optimization are as follows:

[0062] S1. Collect the feedback data of the external operator and label it

[0063] After each rod-changing operation, the external operator provides feedback on the precision compensation effect through the operation interface (such as touch screen, operation buttons, voice input, etc.). These feedback data contain the operator's feelings about the current task execution, such as whether the anchor rod is accurately output, whether it is stuck, whether the pressure is appropriate, whether there are problems such as operation failure, etc. Subsequently, the collected feedback data is labeled, that is, each feedback message is converted into a standard format that the device can understand. Labeling can use different categories according to the actual situation, such as:

[0064] Positive labels: such as "successfully output", "appropriate pressure", "accurate grasping", etc.;

[0065] Negative labels: such as "output failure", "anchor rod stuck", "excessive pressure", or "gripper did not grasp", etc.

[0066] S2. Integrate the labeled data into the training set of the model as calibration data

[0067] The labeled data, together with the real-time sensor data collected by the device (such as the rotation angle of the disc, the pressure of the rubber roller, the speed of the servo cylinder, etc.), is integrated into the training set of the learning model as new training data. These data will help to supplement and correct the deficiencies of the model, especially improve the model's understanding of the device's performance in different working environments;

[0068] Subsequently, the feedback data provided by the external operator is used as calibration data to help the device understand the error patterns that occur in specific situations. The calibration data includes not only the operator's feedback on the success or failure of the task, but also the operator's evaluation of the quality of task completion and the description of specific problems.

[0069] S3. Through the deep reinforcement learning method, the accuracy compensation learning model uses the calibration data for corrective training to complete the optimization.

[0070] After integrating the data of the external operator's feedback, the deep reinforcement learning method is used to optimize the accuracy compensation learning model. Through deep reinforcement learning, the model can adjust its strategy according to the new feedback data, enabling the device to make optimal decisions in the face of various different situations.

[0071] By using the calibration data in the reinforcement learning process, the deep reinforcement learning model will continuously correct its compensation strategy. For example, if a certain operation often deviates under specific environmental conditions, the model will be corrected through the feedback data, adjust the control signal of the servo motor, adjust the pressure of the cylinder, or modify the actions of other key components to adapt to the actual situation.

[0072] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the framework and scope of application of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic rod changing mechanism for a coal mine drilling and anchoring robot, comprising a mounting plate (1), a six-axis mechanical arm (2), a servo-pneumatic gripper (3) and a microcontroller, characterized in that: A servo linear module (4) is installed on the mounting plate (1), and the servo linear module (4) is connected to the six-axis robot (2) through a fixing plate, and a servo pneumatic clamp (3) is installed on the six-axis robot (2). An anchor rod storage cylinder (5) is installed above the mounting plate (1) through a support frame, and one end of the anchor rod storage cylinder (5) is provided with an anchor rod outlet (6), a circular disc (7) is provided on the inner side of the anchor rod storage cylinder (5), and the circular disc (7) is evenly arranged in a plurality of groups, and the plurality of groups of the circular disc (7) are connected to the anchor rod storage cylinder (5) through a driving assembly, and a first opening (8) is provided at the upper end of the anchor rod storage cylinder (5), an L-shaped assembly plate (9) is installed on the mounting plate (1), and an anchor rod output assembly is installed on the L-shaped assembly plate (9), and the position of the anchor rod output assembly corresponds to the first opening (8), and an anchor rod bayonet (10) is provided on the circular disc (7), and the anchor rod bayonet (10) is evenly arranged in a plurality of groups.

2. The automatic rod changing mechanism of a coal mine drilling and anchoring robot according to claim 1, characterized in that: The anchor rod output assembly comprises a servo cylinder (11), wherein the servo cylinder (11) is evenly arranged in a plurality of groups, the output end of the servo cylinder (11) passes through an L-shaped assembly plate (9) and is provided with a fixing frame (12), a rubber roller (13) is provided on the fixing frame (12) via a bearing, and the rubber roller (13) is driven by a wheel hub motor.

3. The automatic rod changing mechanism of a coal mine drilling and anchoring robot according to claim 1, characterized in that: One end of the anchor rod bayonet (10) is arranged in an arc shape, and the other end of the anchor rod bayonet (10) is communicated with the outside. A universal ball (14) is installed inside the anchor rod bayonet (10), and a plurality of groups of universal ball (14) are evenly arranged.

4. The automatic rod changing mechanism of a coal mine drilling and anchoring robot according to claim 1, characterized in that: The driving assembly comprises a main shaft (15), both ends of which are connected to the anchor rod storage tube (5) through bearings, and the main shaft (15) is driven by a servo motor (16). An extension block (17) is installed on the main shaft (15), and a clamping groove (18) is provided on the inner side of the disc (7). The number and position of the clamping groove (18) and the extension block (17) are consistent, and the clamping groove (18) is adapted to the extension block (17).

5. The automatic rod changing mechanism of a coal mine drilling and anchoring robot according to claim 1, characterized in that: An extension rod (19) is mounted on the mounting plate (1), an arc-shaped positioning block (20) is mounted on the upper end of the extension rod (19), and the lower end vertex of the arc-shaped positioning block (20) is on the same horizontal line as the lower end vertex of the anchor rod outlet (6).

6. The automatic rod changing mechanism of a coal mine drilling and anchoring robot according to claim 1, characterized in that: Protective rubber pads (21) are installed at both ends of the inner side of the anchor rod storage tube (5).

7. A precision compensation method for an automatic rod-changing mechanism of a coal mine drilling and anchoring robot applied to claims 1 to 6, characterized in that: The following steps are involved: Step 1: Data Collection By means of arranging sensors, the rotation angle data of the disk (7), the contact pressure data between the rubber roller (13) and the anchor rod, the speed data of the servo linear module (4) and the servo cylinder (11), and the operation position data of the six-axis robot arm (2) are collected; Step 2: Build a precision compensation learning model A precision compensation learning model is constructed based on a random forest regression model and a deep reinforcement learning method, and is trained using the collected sensor data and historical operation data so that the precision compensation learning model can compensate for the detected errors based on the current sensor data; Step 3: Deployment and use of the precision compensation learning model The precision compensation learning model is deployed in the microcontroller, which processes the collected real-time data in real time and outputs the prediction results, and then dynamically adjusts the precision compensation according to the results; Step 4: Feedback and optimization of accuracy compensation Through the intervention of external operators, feedback on the effect of precision compensation is collected and the precision compensation learning model is optimized.

8. The accuracy compensation method of the automatic rod-changing mechanism of a coal mine drilling and anchoring robot according to claim 7 is characterized by: In step 4, the specific steps of optimization are: S1. Collect feedback data from external operators and label them; S2, integrate the labeled data into the training set of the model as calibration data; S3. Through the deep reinforcement learning method, the precision compensation learning model uses the calibration data for correction training to complete the optimization.

Images