A coal mine drilling and anchoring robot automatic rod changing mechanism and a precision compensation method thereof

Abstract

The present application provides a kind of coal mine drill anchor robot automatic rod changing mechanism and its precision compensation method, it is related to coal mine drill anchor robot rod changing technical field, including mounting plate, six-axis robot arm, servo pneumatic gripper and microcontroller, the upper portion of mounting plate is equipped with anchor rod storage cylinder by support frame, and one end of anchor rod storage cylinder is equipped with anchor rod outlet, the inner side of anchor rod storage cylinder is equipped with disc, and disc is uniformly equipped with several groups, several groups of disc are connected with anchor rod storage cylinder by drive assembly, the upper end of anchor rod storage cylinder is equipped with first opening, L-shaped assembly plate is installed on mounting plate, anchor rod output assembly is installed on L-shaped assembly plate, and anchor rod bayonet is equipped on disc;The present application solves the problem of insufficient anchor rod output precision in the existing automatic rod changing system by integrating precision compensation learning model and deep reinforcement learning technology.Through accurate control and real-time error compensation, ensure that anchor rod is smoothly and accurately output, improve the automation level in coal mine operation.

Description

Technical Field

[0001] This invention relates to the field of rod changing technology for coal mine drilling and anchoring robots, and in particular to an automatic rod changing mechanism for coal mine drilling and anchoring robots and its accuracy compensation method. Background Technology

[0002] With the rapid development of science and technology, intelligent coal mining has become an important development direction. In recent years, the pace of automation and intelligentization of fully mechanized mining faces has been accelerating, posing severe challenges to the production capacity and advancement speed of fully mechanized tunneling faces. This necessitates a corresponding increase in the speed of mining, tunneling, and support operations. Currently, most coal mines in China have partially mechanized mining, tunneling, and support operations, using tunneling machines for roadway excavation and drilling and anchoring machinery or integrated drilling and anchoring machines for support. These operations are largely conducted with personnel following the machines, but the drilling and anchoring speed lags significantly behind the tunneling speed, resulting in poor working conditions, high labor intensity, and low tunneling efficiency. This fails to guarantee the progress requirements of fully mechanized mining, directly impacting the safe, high-yield, and efficient mining of coal.

[0003] Coal mine drilling and anchoring robots, as integrated automated equipment, are widely used in coal mining, tunnel engineering, and underground structure reinforcement. Their main function is to automatically complete drilling, anchor bolt installation, and support work, reducing manual operation and improving work efficiency and safety. Coal mine drilling and anchoring robots typically consist of core components such as a robotic arm, servo motors, a control system, and sensors, enabling them to perform high-precision drilling and anchor bolt installation tasks.

[0004] In the application of coal mine drilling and anchoring robots, automatic rod changing systems are an important component for improving work efficiency, reducing manual intervention, and enhancing operational safety. Coal mine drilling and anchoring robots can automatically complete drilling and anchor installation tasks, greatly improving the efficiency and safety of coal mine support operations. However, in existing automatic rod changing technologies, the accuracy of anchor output typically relies 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 anchors not being accurately output from the anchor storage cylinder to the target position. Because the system cannot adjust the output process based on real-time data, situations often arise where anchors get stuck, misaligned, or fail to be smoothly delivered to the robotic arm's gripping area. Furthermore, the lack of a real-time accuracy compensation mechanism during anchor output leads to inaccurate anchor output positions, thus affecting the accuracy of subsequent operations. Therefore, this invention proposes an automatic rod changing mechanism for coal mine drilling and anchoring robots and its accuracy compensation method to solve the problems existing in the prior art. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide an automatic rod-changing mechanism and its accuracy compensation method for a coal mine drilling and anchoring robot. This automatic rod-changing mechanism and its accuracy compensation method for a coal mine drilling and anchoring robot have the advantage of improving the output accuracy of anchor bolts during the automatic rod-changing process and can solve the problems existing in the prior art.

[0006] To achieve the objectives of this invention, the following technical solution is provided: an automatic rod changing mechanism for a coal mine drilling and anchoring robot, comprising a mounting plate, a six-axis robotic arm, a servo pneumatic gripper, and a microcontroller. A servo linear module is mounted on the mounting plate and connected to the six-axis robotic arm via a fixing plate. A servo pneumatic gripper is mounted on the six-axis robotic arm. An anchor rod storage cylinder is mounted above the mounting plate via a support frame, and one end of the anchor rod storage cylinder has an anchor rod outlet. A disc is provided inside the anchor rod storage cylinder, and several groups of discs are evenly distributed. These discs are connected to the anchor rod storage cylinder via a drive assembly. A first opening is provided at the upper end of the anchor rod storage cylinder. An L-shaped assembly plate is mounted on the mounting plate, and an anchor rod output assembly is mounted on the L-shaped assembly plate, with the position of the anchor rod output assembly corresponding to the first opening. An anchor rod latches are provided on the discs, and several groups of anchor rod latches are evenly distributed.

[0007] A further improvement is that the anchor output assembly includes a servo cylinder, and several sets of the servo cylinder are evenly arranged. The output end of the servo cylinder passes through an L-shaped mounting plate and is mounted on a fixing frame. A rubber roller is mounted on the fixing frame via a bearing, and the rubber roller is driven by a hub motor.

[0008] A further improvement is that one end of the anchor bolt clamp is arc-shaped, and the other end of the anchor bolt clamp is connected to the outside. Universal ball bearings are installed on the inner side of the anchor bolt clamp, and several sets of universal ball bearings are evenly arranged.

[0009] A further improvement is that the drive assembly includes a main shaft, both ends of which are connected to the anchor rod storage cylinder via bearings, and the main shaft is driven by a servo motor. An extension block is installed on the main shaft, and a slot is provided on the inner side of the disc. The number and position of the slots and the extension blocks are the same, and the slots are adapted to the extension blocks.

[0010] A further improvement is 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 is at the same horizontal line as the lower vertex of the anchor bolt outlet.

[0011] A further improvement is that protective rubber pads are installed at both ends of the inner side of the anchor rod storage cylinder.

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

[0013] Step 1: Data Collection

[0014] By deploying sensors, data on the rotation angle of the disc, the contact pressure between the rubber roller and the anchor rod, the speed data of the servo linear module and the servo cylinder, and the operating position data of the six-axis robotic arm are collected.

[0015] Step 2: Construct a precision compensation learning model

[0016] A precision compensation learning model is constructed based on a random forest regression model and a deep reinforcement learning method. It is trained using collected sensor data and historical operation data, enabling the precision compensation learning model to compensate for detected errors based on current sensor data.

[0017] Step 3: Deployment and Use of the Precision Compensation Learning Model

[0018] The accuracy compensation learning model is deployed in a microcontroller, which processes the collected real-time data in real time and outputs the prediction results. The accuracy compensation is then dynamically adjusted based on the results.

[0019] Step 4: Feedback and Optimization of Accuracy Compensation

[0020] By involving external operators, feedback on the accuracy compensation effect is collected, and the accuracy compensation learning model is optimized.

[0021] The further improvement lies in the following: In step four, the specific optimization steps are as follows:

[0022] S1. Collect feedback data from external operators and label it;

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

[0024] S3. Through deep reinforcement learning, the accuracy compensation learning model is trained and corrected using calibration data to achieve optimization.

[0025] The beneficial effects of this invention are as follows:

[0026] (1) This device integrates key components such as servo cylinders, rubber rollers, and disc rotation control, and combines them with a precision compensation learning model to achieve precise control of the anchor bolt output accuracy. In particular, during anchor bolt output, it can ensure that the anchor bolt accurately passes through the first opening and smoothly enters the gripping area of ​​the robotic arm each time through real-time sensor feedback and precision compensation mechanism, avoiding problems such as anchor bolt output failure or jamming.

[0027] (2) This device employs deep reinforcement learning and random forest regression models, enabling real-time compensation for errors during the anchor bolt output process. By collecting data such as the disk rotation angle and rubber roller pressure in real time, it can dynamically adjust the motion parameters of components such as the servo motor and cylinder to compensate for the effects of mechanical errors and environmental changes, ensuring the accuracy and stability of the anchor bolt output process. Simultaneously, it can automatically optimize the control strategy based on feedback data after each operation, thereby improving accuracy and operational stability. Attached Figure Description

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

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

[0030] Figure 3 This is a side view of the anchor storage cylinder structure of the present invention.

[0031] Figure 4 This is a front view of the structure after the anchor rod of the present invention has been removed.

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

[0033] The components include: 1. Mounting plate; 2. Six-axis robotic arm; 3. Servo pneumatic gripper; 4. Servo linear module; 5. Anchor bolt storage cylinder; 6. Anchor bolt outlet; 7. Disc; 8. First opening; 9. L-shaped assembly plate; 10. Anchor bolt bayonet; 11. Servo cylinder; 12. Fixing frame; 13. Rubber roller; 14. Universal ball bearing; 15. Spindle; 16. Servo motor; 17. Extension block; 18. Slot; 19. Extension rod; 20. Arc-shaped positioning block; 21. Protective rubber pad. Detailed Implementation

[0034] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0035] Traditional automatic anchor bolt changing systems lack real-time accuracy compensation mechanisms during anchor bolt output. Driven devices such as cylinders and servo motors are prone to performance deviations after prolonged operation, and existing systems cannot dynamically adjust based on sensor feedback. This results in inaccurate anchor bolt output positions, affecting the accuracy of subsequent operations. Especially in complex working environments, external factors such as equipment wear and sensor drift often lead to reduced output accuracy.

[0036] Meanwhile, in existing technologies, errors are prone to occur during the movement of mechanical components (such as discs, cylinders, and grippers). These errors may accumulate over time, affecting the output accuracy of the anchor bolt. Especially when multiple mechanical components work together, even minor deviations may be amplified, leading to an increase in the overall output error, thereby affecting the stability and efficiency of the entire bolt replacement process.

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

[0038] First, the six-axis robotic arm 2 positions itself at the drilling location of the coal mine drilling and anchoring robot. Then, a servo-driven pneumatic gripper 3 grasps the old anchor rod in the current borehole, pulls it out, and transfers it to the corresponding abandoned anchor rod storage area. Next, a new anchor rod is retrieved from the anchor rod storage cylinder 5, and the six-axis robotic arm 2 installs it at the current borehole position, thus completing an automatic anchor rod replacement process. During the accuracy compensation process, this device mainly involves the positioning accuracy of the new anchor rod output, the contact accuracy of the new anchor rod delivery, and the gripping 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 with all the electronic components in this device and then controlling them.

[0040] A servo linear module 4 is mounted on mounting plate 1, and the servo linear module 4 is connected to the six-axis robotic arm 2 via a fixing plate. A servo pneumatic gripper 3 is mounted on the six-axis robotic arm 2. The servo linear module 4 consists of a high-precision servo motor and precision guide rails, and its function is to provide precise linear motion. Its sliding end is fixedly connected to the base of the six-axis robotic arm 2, meeting 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 of the six-axis robotic arm 2, ensuring that the gripper can accurately perform grasping and releasing operations as the six-axis robotic arm 2 moves. Mounting plate 1 is used to fix the robot to the coal mine drilling and anchoring robot.

[0041] An anchor rod storage cylinder 5 is mounted on top of the mounting plate 1 via a support frame. One end of the anchor rod storage cylinder 5 is provided with an anchor rod outlet 6 (right end in this embodiment). In this embodiment, the anchor rod outlet 6 can only accommodate one set of anchor rods entering and exiting, ensuring stable output of anchor rods and precise gripping by the six-axis robotic arm 2. Correspondingly, a disc 7 is provided on the inner side of the anchor rod storage cylinder 5, and several sets of discs 7 are evenly distributed. The function of the discs 7 is to separate the anchor rods in the anchor rod storage cylinder 5. In this embodiment, five sets of discs 7 are evenly distributed. At the same time, an anchor rod slot 10 is provided on the disc 7, and several sets of anchor rod slots 10 are evenly distributed. The diameter of the disc 7 is smaller than the diameter of the inner cavity of the anchor rod storage cylinder 5. Thus, the design of the disc 7 allows each set of anchor rods to be stably stored in the anchor rod slot 10 of each disc 7, avoiding instability of the anchor rods due to excessive space. Several sets of discs 7 are connected to the anchor rod storage cylinder 5 via a drive assembly, which includes a main shaft 15. Both ends of the main shaft 15 are connected to the anchor rod storage cylinder 5 via bearings, and the main shaft 15 is driven by a servo motor 16 located on the outside of the anchor rod storage cylinder 5 (away from the end of the anchor rod outlet 6). An extension block 17 is installed on the main shaft 15, and the extension block 17 is integrally formed with the main shaft 15. The inner side of the disc 7 is provided with a slot 18. The number and position of the slots 18 and the extension blocks 17 are the same, and the slots 18 are adapted to the extension blocks 17. In this embodiment, the inner side of each set of discs 7 is provided with six sets of slots 18, and each set of slots 18 is provided with an extension block 17. Furthermore, the extension blocks 17 are fixedly connected to the discs 7. By setting the extension blocks 17, the contact area between the main shaft 15 and the discs 7 is increased, thereby improving its stability and ensuring the positional 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 anchor bolt storage cylinder 5 has a first opening 8. An L-shaped assembly plate 9 is installed on the mounting plate 1, with one end of the L-shaped assembly plate 9 extending directly above the anchor bolt storage cylinder 5. An anchor bolt output assembly is installed on the L-shaped assembly plate 9, and the position of the anchor bolt output assembly corresponds to the first opening 8. The anchor bolt output assembly includes a servo cylinder 11, and several sets 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 mounted on a fixing frame 12. A rubber roller 13 is mounted on the fixing frame 12 via bearings, and the rubber roller 13 is driven by a hub motor. The hub motor provides power through motor control, driving the rubber roller 13 to rotate, thereby generating friction and pushing the anchor bolt along a designated path to the anchor bolt outlet 6. Furthermore, during operation, the speed and torque of the motor precisely control the speed of the rubber roller to ensure that the anchor bolt can be output smoothly at a suitable speed. In this device, each set of anchor rods is delivered. A rotating disc 7 rotates the anchor rod to be delivered directly below the first opening 8. Then, a servo cylinder 11 drives the fixed frame 12 downwards, causing the rubber roller 13 to contact the anchor rod (the position and contact pressure of the rubber roller 13 can be precisely controlled by adjusting the output force and stroke of the servo cylinder 11). In this device, one end of the anchor rod bayonet 10 is arc-shaped, and the other end is connected to the outside. Universal ball bearings 14 are installed inside the anchor rod bayonet 10, and several sets of universal ball bearings 14 are evenly distributed. The function of the universal ball bearings 14 is to facilitate the movement of the anchor rod. Then, the rubber roller 13 starts, and through the friction generated by the contact, it drives the anchor rod to move out of the anchor rod outlet 6. At this time, the six-axis robotic arm 2 drives the servo pneumatic gripper 3 into the predetermined gripping area to grip the moved anchor rod. The servo pneumatic gripper 3 can adjust the gripping force through pneumatic control to ensure safety and stability during the gripping process.

[0043] Correspondingly, this device achieves a precise anchor bolt output process through the coordinated operation of key components such as the anchor bolt storage cylinder 5, the disc 7, the servo cylinder 11, and the rubber roller 13. Each output anchor bolt is moved to below the first opening 8 by the rotating disc 7, and then 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. 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 at the same horizontal line as the lower vertex of the anchor rod outlet 6. When the anchor rod moves a certain distance, it contacts the arc-shaped positioning block 20, which plays a role in precise guidance.

[0045] Furthermore, protective rubber pads 21 are installed at both ends of the inner side of the anchor bolt storage cylinder 5 to prevent the anchor bolt from directly colliding with the two ends of the anchor bolt storage cylinder 5.

[0046] A method for accuracy compensation of an automatic rod-changing mechanism in a coal mine drilling and anchoring robot includes the following steps:

[0047] Step 1: Data Acquisition (Data is collected from the entire lever-changing mechanism by deploying sensors to obtain key parameters and real-time feedback, thereby enabling accurate precision compensation)

[0048] By deploying sensors, data on the rotation angle of the disc 7, the contact pressure 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 operating position data of the six-axis robotic arm 2 are collected.

[0049] Specifically, a rotary encoder is installed on disc 7 to monitor the disc's rotation angle in real time. The rotary encoder provides high-precision angle data, ensuring that the disc's rotation process is error-free, thus guaranteeing that the anchor rod can accurately reach the predetermined position each time it is replaced. This data is used to determine the current rotation state of the disc, ensuring that the rotation angle of disc 7 has been adjusted to the correct position each time the anchor rod is output, preparing for subsequent operations.

[0050] A pressure sensor (such as a strain gauge pressure sensor or a piezoelectric sensor) is installed in the area where the rubber roller 13 contacts the anchor bolt to monitor the contact pressure between the two in real time. The pressure sensor can detect changes in contact force and adjust the roller pressure through feedback data to avoid affecting the delivery of the anchor bolt due to excessive or insufficient pressure. This data is used to understand the frictional force between the roller and the anchor bolt, ensuring that the pressure is always controlled within the optimal range during the anchor bolt output process, thereby avoiding anchor bolt damage or slippage.

[0051] Speed ​​sensors or encoders are installed on the servo linear module 4 and the servo cylinder 11 to monitor the speed and motion status of these two drive devices in real time. These sensors can provide linear displacement data and speed data, helping the control device to 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 actuators move along a predetermined path, avoiding errors or delays.

[0052] A vision sensor is installed on the six-axis robotic arm 2 to acquire the robotic arm's operational position data in real time. The position sensor can provide the precise position and attitude of the robotic arm's end effector, ensuring the positioning accuracy when gripping the anchor rod. It is used to precisely control the robotic arm's gripping action, ensuring that the robotic arm can accurately dock at each critical step in the anchor rod delivery process, and avoiding positional errors when performing the gripping task.

[0053] Step 2: Construct a precision compensation learning model

[0054] A precision compensation learning model is constructed based on a random forest regression model and a deep reinforcement learning method. It is trained using collected sensor data and historical operation data, enabling the precision compensation learning model to compensate for detected errors based on current sensor data.

[0055] Random forest regression is an ensemble learning method that generates multiple decision trees for prediction and outputs the prediction result as the average of the trees. In the accuracy compensation model, the application of random forest regression can help predict the possible errors of various components (such as disks, cylinders, and robotic arms) under given conditions. Correspondingly, using the sensor data collected in step one (such as disk rotation angle, pressure, speed, and robotic arm position) as input features, the random forest regression model can predict the possible errors of the device based on 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 training, historical operating data and known error values ​​are used to train the random forest model. The historical data should contain error information under various operating conditions (e.g., the possible errors of the device when the ambient temperature changes or the sensor fails). The trained model can predict the accuracy error of the device based on real-time input data and provide corresponding compensation suggestions.

[0056] Deep reinforcement learning is a technique that combines deep learning and reinforcement learning. It optimizes policies through interaction with the environment. During training, the agent operates the environment using a simulated device, learning the optimal policy with each interaction and optimizing compensation decisions through a reward function. Through repeated trial and error, the agent gradually learns how to reduce errors and improve accuracy. Accordingly, by combining random forest regression models and deep reinforcement learning, an efficient accuracy compensation learning model can be constructed. The random forest regression model is used to predict errors, while the deep reinforcement learning method adaptively adjusts and optimizes the compensation strategy. This enables precise error compensation based on real-time sensor data, improving accuracy and stability during lever changes. Furthermore, it can self-learn and optimize in a constantly changing working environment, thus providing long-term, continuous accuracy assurance.

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

[0058] The accuracy compensation learning model is deployed within a microcontroller. The microcontroller processes the collected real-time data and outputs prediction results, which are then dynamically adjusted to improve accuracy compensation. The microcontroller uses an ARM Cortex series processor, which possesses strong computing power and high-speed data processing capabilities. Therefore, embedding the accuracy compensation learning model into the microcontroller allows for real-time processing of sensor data, generation of compensation outputs, and final control.

[0059] Correspondingly, the microcontroller receives feedback data from various sensors in real time, performs necessary preprocessing (denoising, normalization, and smoothing) on ​​this data, and then inputs it into the model. Based on the model's prediction results, the microcontroller dynamically adjusts the control parameters of the device. Specifically, the control device uses actuators such as the servo motor 16, servo cylinder 11, and six-axis robotic arm 2 to perform precision compensation adjustments to ensure that the motion, pressure, speed, and other parameters of the device remain in optimal condition during the automatic rod changing process. 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 high or too low, 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 uneven delivery.

[0060] Step 4: Feedback and Optimization of Accuracy Compensation

[0061] By involving external operators, feedback on the accuracy compensation effect is collected, and the accuracy compensation learning model is optimized. The specific optimization steps are as follows:

[0062] S1. Collect feedback data from external operators and label it.

[0063] After each boom change, external operators provide feedback on the accuracy compensation effect via an interface (such as a touchscreen, operation buttons, or voice input). This feedback data includes the operator's perception of the current task execution, such as whether the boom is output accurately, whether it is stuck, whether the pressure is appropriate, and whether there are any operational failures. The collected feedback data is then tagged, that is, each piece of feedback information is converted into a standard format that the device can understand. Tagging can use different categories depending on the actual situation, such as:

[0064] Positive labels: such as "successful output", "moderate pressure", "accurate capture", etc.

[0065] Negative labels: such as "output failed", "anchor stuck", "excessive pressure" or "gripper not gripped", etc.

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

[0067] The labeled data, along with real-time sensor data collected by the device (such as disk rotation angle, rubber roller pressure, servo cylinder speed, etc.), are integrated into the training set of the learning model as new training data. This data will help supplement and correct the model's shortcomings, especially improving the model's understanding of the device's performance under different operating environments;

[0068] The feedback data provided by external operators then serves as calibration data, helping the device understand error patterns that occur in specific situations. The calibration data includes not only operator feedback on whether the task was successful or not, but also the operator's evaluation of the quality of task completion and descriptions of specific problems.

[0069] S3. Through deep reinforcement learning, the accuracy compensation learning model uses calibration data for correction and training, thus completing the optimization.

[0070] After integrating data from external operator feedback, deep reinforcement learning methods are used to optimize the accuracy compensation learning model. Through deep reinforcement learning, the model can adjust its strategy based on new feedback data, enabling the device to make optimal decisions in various situations.

[0071] By using calibration data during reinforcement learning, deep reinforcement learning models continuously refine their compensation strategies. For example, if a certain operation frequently deviates under specific environmental conditions, the model will correct itself using feedback data, adjusting the control signal of the servo motor, adjusting the pressure of the cylinder, or modifying the actions of other key components to adapt to the actual situation.

[0072] The foregoing has shown and described 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 to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the present invention without departing from its framework and scope of application, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of 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 robotic arm (2), a servo pneumatic gripper (3), and a microcontroller, characterized in that: A servo linear module (4) is mounted on the mounting plate (1), and the servo linear module (4) is connected to a six-axis robotic arm (2) via a fixing plate. A servo pneumatic gripper (3) is mounted on the six-axis robotic arm (2). An anchor rod storage cylinder (5) is mounted on the top of the mounting plate (1) via a support frame, and an anchor rod outlet (6) is provided at one end of the anchor rod storage cylinder (5). A disc (7) is provided on the inner side of the anchor rod storage cylinder (5), and several groups of discs (7) are evenly arranged. Several groups of discs (7) are connected to the anchor rod storage cylinder (5) via a drive assembly. A first opening (8) is provided at the upper end of the anchor rod storage cylinder (5). An L-shaped assembly plate (9) is mounted on the mounting plate (1), and an anchor rod output assembly is mounted on the L-shaped assembly plate (9). The position of the anchor bolt output component corresponds to the first opening (8). The disc (7) is provided with anchor bolt slots (10), and several sets of anchor bolt slots (10) are evenly provided. The anchor bolt output component includes a servo cylinder (11), and several sets of servo cylinders (11) are evenly provided. The output end of the servo cylinder (11) passes through the L-shaped assembly plate (9) and is mounted on a fixing frame (12). A rubber roller (13) is mounted on the fixing frame (12) through a bearing, and the rubber roller (13) is driven by a hub motor. One end of the anchor bolt slot (10) is arc-shaped, and the other end of the anchor bolt slot (10) is connected to the outside. A universal ball bearing (14) is installed on the inner side of the anchor bolt slot (10), and several sets of universal ball bearings (14) are evenly provided.

2. The automatic rod-changing mechanism for a coal mine drilling and anchoring robot according to claim 1, characterized in that: The drive assembly includes a main shaft (15), both ends of which are connected to the anchor rod storage cylinder (5) via bearings. The main shaft (15) is driven by a servo motor (16). An extension block (17) is mounted on the main shaft (15). A slot (18) is provided on the inner side of the disc (7). The number and position of the slot (18) and the extension block (17) are the same, and the slot (18) is adapted to the extension block (17).

3. The automatic rod-changing mechanism for a coal mine drilling and anchoring robot according to claim 1, characterized in that: An extension rod (19) is installed on the mounting plate (1). 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 anchor bolt outlet (6).

4. The automatic rod-changing mechanism for 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 storage tube (5).

5. A method for accuracy compensation of an automatic rod-changing mechanism for a coal mine drilling and anchoring robot, characterized in that: The accuracy compensation method for the automatic rod changing mechanism of the coal mine drilling and anchoring robot is applied to the automatic rod changing mechanism of the coal mine drilling and anchoring robot as described in any one of claims 1-4. The accuracy compensation method for the automatic rod changing mechanism of the coal mine drilling and anchoring robot includes the following steps: Step 1: Data Collection By deploying sensors, 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) are collected. Step 2: Construct 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. It is trained using collected sensor data and historical operation data, enabling the precision compensation learning model to compensate for detected errors based on current sensor data. Step 3: Deployment and Use of the Precision Compensation Learning Model The accuracy compensation learning model is deployed in a microcontroller, which processes the collected real-time data in real time and outputs the prediction results. The accuracy compensation is then dynamically adjusted based on the results. Step 4: Feedback and Optimization of Accuracy Compensation By involving external operators, feedback on the accuracy compensation effect is collected, and the accuracy compensation learning model is optimized.

6. The accuracy compensation method for an automatic rod-changing mechanism of a coal mine drilling and anchoring robot according to claim 5, characterized in that: In step four, the specific steps for optimization are as follows: S1. Collect feedback data from external operators and label it; S2. Integrate the labeled data into the model's training set as calibration data; S3. Through deep reinforcement learning, the accuracy compensation learning model is trained and corrected using calibration data to achieve optimization.

Images