Control method and system for hydraulic support

Abstract

This invention discloses a control method and system for hydraulic supports, belonging to the field of advanced support technology for coal mine roadways. The control method and system utilize roadway environmental characteristics to initialize and calibrate the pose detection module, providing a precise starting reference for the inertial navigation system (INS) to solve the initial alignment problem in the absence of GPS underground. During the support relocation process, the system senses the roadway environment in real time and extracts features to correct the cumulative drift caused by the INS over long distances, ensuring that the positioning accuracy continuously meets the support spacing requirements. Finally, the system automatically stops and supports the system based on the corrected coordinates. Each step is tightly coupled through intermediate variables such as the starting reference, roadway environmental characteristics, and corrected positioning coordinates, achieving physical isolation between operators and hazardous areas and fully automated support relocation throughout the entire process. This effectively solves the problems of high personnel safety risks, reliance on manual experience for positioning, and the inability to self-correct INS drift in existing technologies, providing a safe and efficient control solution for advanced support in coal mine roadways.

Description

Technical Field

[0001] This invention relates to the field of advanced support technology for coal mine roadways, specifically to a control method and system for hydraulic supports. Background Technology

[0002] Currently, in the construction of coal mine roadways, longwall mining roadways, and large chambers, temporary support for the roadway roof is usually required to ensure its stability. However, most existing pre-support equipment uses single hydraulic props in conjunction with articulated roof beams, stepping pre-supports, or support equipment that relies on manual towing. In actual use, this type of equipment requires frequent cycles of "support-unloading-moving-re-supporting". As the support base and roof beam continuously apply concentrated loads to the roof during repeated support, it is easy to disturb and damage the original anchor bolt and cable support system in the roadway. Especially in deep mines or under fractured roof conditions, repeated stress can lead to a decrease in anchor bolt preload, instability of anchor cables, and expansion of surrounding rock fissures, further damaging the original load-bearing structure of the roof. In severe cases, it can even cause safety accidents such as roof delamination, roof fall, and local collapse.

[0003] In addition, traditional advanced support equipment generally suffers from low mobility, high manual involvement, high operational risks, and difficulty in adapting to complex roadway environments. Especially in long roadways, the movement of supports often requires close-range operation by personnel. In the presence of cables, liquid pipes, belt conveyors, and transfer equipment, collisions, squeezing, and accidental injuries to personnel are prone to occur. Therefore, it is no longer able to meet the development needs of intelligent and unmanned coal mines.

[0004] Currently, in the process of advanced support for coal mine roadways, tracked hydraulic supports are beginning to be used to replace traditional single hydraulic props or towed advanced supports, in order to improve the mobility of the support equipment and its ability to adapt to complex floor environments. Existing tracked hydraulic supports typically consist of a tracked walking mechanism, hydraulic props, a top beam, hydraulic valves, and a manual operation control system. The tracked mechanism drives the support to move along the roadway, while the hydraulic props control the raising and lowering of the top beam, thereby achieving roof support and release. After completing the support of the current area, the equipment needs to be manually controlled to move the tracks forward to a new support area, and then the hydraulic props are raised again to complete the roof support. Current systems mainly employ a manual, close-range operation method. Operators typically control the tracks forward, backward, and turning, as well as the prop movement, through hydraulic operating valves on the equipment itself, a handheld wired controller, or a simple wireless remote control. Some equipment has basic wireless remote control capabilities, but its control logic still mainly relies on manual observation and experience-based judgment, and a complete intelligent environmental perception and automatic control system has not yet been formed. Therefore, it still belongs to the traditional semi-mechanized support method.

[0005] While existing tracked hydraulic supports offer improved mobility compared to traditional support equipment, they still have several shortcomings. First, most existing equipment relies on close-range manual control, requiring operators to follow the supports for extended periods. In narrow tunnels, complex roof conditions, and environments with cables, hydraulic lines, transfer conveyors, and belt conveyors, collisions, crushing accidents, and falling debris injuries are common, failing to effectively guarantee personnel safety. Second, existing systems lack real-time environmental awareness, typically lacking millimeter-wave radar, visual recognition, or personnel detection capabilities. Therefore, they cannot automatically identify obstacles, personnel, or tunnel boundaries, resulting in low safety during operation. Furthermore, existing tracked supports primarily rely on manual experience for positioning and operation, leading to slow movement and difficulty in achieving precise stopping and rapid support. This can cause a decline in support efficiency during long-distance tunnel operations. In addition, existing technology cannot achieve autonomous obstacle avoidance, path planning, or unmanned collaborative control, failing to meet the development needs of intelligent, reduced-staff, and unmanned coal mines, and cannot provide effective support for future smart mine construction.

[0006] This section is intended to provide background or context for the embodiments of the invention set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section. Summary of the Invention

[0007] In order to solve at least one of the above-mentioned problems in the prior art, embodiments of the present invention provide a control method and system for a hydraulic support.

[0008] This invention provides a control method for a hydraulic support, the hydraulic support including a traveling mechanism, a top beam, a hydraulic column, and a sensing system, the sensing system including a pose detection module and an environmental sensing module, the method including the following steps: The system acquires the tunnel environment information collected by the environmental perception module, extracts tunnel features based on the tunnel environment information, and initializes and calibrates the initial attitude data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support. The initial pose information is then used as the starting reference for the pose detection module to perform dead reckoning. Based on the initial pose information, the hydraulic column is controlled to release the support of the roof, and the traveling mechanism is driven to move the hydraulic support along the roadway. During the movement, the environmental perception module collects tunnel environmental perception data in real time and extracts tunnel environmental features based on the tunnel environmental perception data. The pose detection module performs dead reckoning starting from the initial reference to obtain the dead reckoning position; the cumulative error of the dead reckoning position is corrected using the tunnel environment features to obtain the corrected positioning coordinates. The hydraulic support is stopped according to the corrected positioning coordinates, and the hydraulic column is raised after stopping to complete the support of the top plate.

[0009] In some embodiments, the tunnel environment information includes any one or more of the following: a preset tunnel marker image, a tunnel sidewall outline, a tunnel cross-sectional shape, or fixed equipment features; the tunnel features include any one or more of the following: visual features, geometric contour features, or radar reflection features.

[0010] In some embodiments, the initialization calibration of the initial pose data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support includes: Determine the absolute coordinates of at least one reference target in the tunnel features in the tunnel coordinate system; Obtain the initial pose data detected by the pose detection module; Based on the absolute coordinates of the reference target and the initial attitude data detected by the pose detection module, the initial position coordinates and initial heading angle of the hydraulic support in the roadway coordinate system are calculated, and the initial position coordinates and initial heading angle are used as the initial pose information.

[0011] In some embodiments, the method further includes: During the movement, safety assessments are made in real time based on the environmental perception data of the tunnel collected by the environmental perception module. Based on the safety assessment results, the hydraulic support is controlled to execute corresponding obstacle avoidance strategies to ensure the safety of the support operation.

[0012] In some embodiments, the environmental perception module includes a ranging sensor, a camera, and an infrared detection device.

[0013] In some embodiments, the ranging sensor is a millimeter-wave radar, which is deployed in at least four directions: front, rear, left, and right of the hydraulic support; the camera is deployed in at least four directions: front, rear, left, and right of the hydraulic support; and the infrared detection device is deployed in at least the front and rear directions of the hydraulic support.

[0014] In some embodiments, the safety determination based on the roadway environmental perception data collected by the environmental perception module includes: Visual images are simultaneously acquired by the camera installed on the hydraulic support, infrared thermal images are simultaneously acquired by the infrared detection device, and the surrounding environment is continuously scanned by the millimeter-wave radar to obtain millimeter-wave radar ranging data. Obstacle features are determined based on the visual image, and the obstacle features are corrected based on the millimeter-wave radar ranging data to obtain the true physical features of the obstacle; Based on the infrared thermal imaging and the visual images, an intrusion determination of personnel into the dangerous area is made; A safety assessment is made based on the actual physical characteristics of the obstacle and the results of the personnel intrusion assessment of the dangerous area.

[0015] In some embodiments, determining obstacle features based on the visual image includes: Lens distortion correction is performed on the visual images captured by each camera based on the camera intrinsic parameters of each camera to obtain the corrected visual image. Feature points of the overlapping region of the corrected visual images corresponding to adjacent cameras are extracted using a feature matching algorithm. The homography matrix is ​​solved based on the feature points. The corrected visual images are then registered based on the homography matrix to obtain the image to be stitched together. A pixel-weighted fusion algorithm is used to stitch the images to be stitched together to eliminate stitching gaps and brightness abrupt changes, generating a 360° panoramic view image of the hydraulic support. Obstacle features in the 360° panoramic view image are extracted using an edge detection algorithm.

[0016] In some embodiments, correcting the obstacle features based on the millimeter-wave radar ranging data to obtain the true physical features of the obstacle includes: Based on the millimeter-wave radar ranging data and pixel spatial coordinate mapping algorithm, and according to the camera's internal and external parameter calibration parameters, the obstacle features are transformed from pixel coordinate system to underground physical coordinate system to obtain the true physical features of the obstacle.

[0017] In some embodiments, determining whether personnel have intruded into a dangerous area based on the infrared thermal image and the visual image includes: Based on a temperature threshold, pixels in the infrared thermal image are filtered to obtain suspected human targets in the infrared thermal image. The suspected human targets are screened using a connected component morphological feature discrimination mechanism to obtain the human target screening results; The human target recognition result is obtained by identifying human targets in the danger zone of the 360° panoramic view image using machine vision algorithms. Based on the results of human target screening and human target recognition, a determination of human intrusion into dangerous areas is made.

[0018] In some embodiments, the step of performing dead reckoning using the pose detection module with the starting reference as the starting point to obtain the dead reckoning position includes: The initial position coordinates in the initial pose information are used as the integration starting point, and the initial heading angle in the initial pose information is used as the initial heading. The posture detection module collects the acceleration data of the hydraulic support along the roadway direction in real time. The acceleration data is integrated twice to obtain the displacement increment of the hydraulic support relative to the initial reference. The displacement increment is then added to the initial position coordinates to obtain the real-time calculated position coordinates.

[0019] In some embodiments, correcting the cumulative error of the dead reckoning position using the tunnel environment features to obtain the corrected positioning coordinates includes: Based on the geometric contour features and / or spatial distribution features in the roadway environment, and combined with the distance data between the hydraulic support and the roadway sidewall collected in real time by the ranging sensor, the first absolute position of the hydraulic support in the transverse direction of the roadway is calculated. Based on the visual features in the tunnel environment, the preset support mileage scale or spacing mark in the tunnel is identified by the environmental image captured by the camera, and the second absolute position of the hydraulic support in the longitudinal direction of the tunnel is obtained. The corrected lateral positioning coordinates are obtained by weighted and fused with the first absolute position using the lateral coordinates of the dead reckoning position. The longitudinal coordinates of the dead reckoning position are weighted and fused with the second absolute position to obtain the corrected longitudinal positioning coordinates; The corrected positioning coordinates are formed by combining the corrected lateral positioning coordinates and the corrected longitudinal positioning coordinates.

[0020] This invention also provides a control system for a hydraulic support, the hydraulic support including a traveling mechanism, a top beam, and a hydraulic column, the control system including: The sensing device includes a pose detection module and an environmental sensing module. The environmental sensing module is used to collect roadway environmental information and to collect roadway environmental sensing data during the movement of the hydraulic support. The pose detection module is used to detect the attitude data of the hydraulic support and perform dead reckoning. The controller, which is signal-connected to the sensing device, the walking mechanism, and the hydraulic column, is configured to perform the following steps: The system acquires the tunnel environment information collected by the environmental perception module, extracts tunnel features based on the tunnel environment information, and initializes and calibrates the initial attitude data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support. The initial pose information is then used as the starting reference for the pose detection module to perform dead reckoning. Based on the initial pose information, the hydraulic column is controlled to release the support of the roof, and the traveling mechanism is driven to move the hydraulic support along the roadway. During the movement, the environmental perception module collects tunnel environmental perception data in real time and extracts tunnel environmental features based on the tunnel environmental perception data. The pose detection module performs dead reckoning starting from the initial reference to obtain the dead reckoning position; the cumulative error of the dead reckoning position is corrected using the tunnel environment features to obtain the corrected positioning coordinates. The hydraulic support is stopped according to the corrected positioning coordinates, and the hydraulic column is raised after stopping to complete the support of the top plate.

[0021] This invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the program to implement the method described in any of the above embodiments.

[0022] This invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the above embodiments.

[0023] This invention also provides a computer program that, when executed by a processor, implements the methods described in any of the above embodiments.

[0024] The hydraulic support control method and system proposed in this invention utilizes roadway environmental characteristics to initialize and calibrate the pose detection module, providing a precise starting reference for the inertial navigation system (INS) to solve the initial alignment problem in the absence of GPS underground. During the support relocation process, the system senses the roadway environment in real time and extracts features to correct the cumulative drift caused by the INS' long-distance operation, ensuring that the positioning accuracy continuously meets the support spacing requirements. Finally, the system automatically stops and supports the system based on the corrected coordinates. Each step is tightly coupled through intermediate variables such as "starting reference," "roadway environmental characteristics," and "corrected positioning coordinates," achieving physical isolation between operators and hazardous areas and fully automated support relocation throughout the entire process. This effectively solves the core problems of existing technologies, such as high personnel safety risks, reliance on manual experience for positioning, and the inability to self-correct INS drift, providing a safe, efficient, and intelligent control solution for advanced support in coal mine roadways. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0026] Figure 2 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0027] Figure 3 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0028] Figure 4 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0029] Figure 5 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0030] Figure 6 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0031] Figure 7 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0032] Figure 8 This is a partial flowchart illustrating a control method for a hydraulic support provided in an embodiment of the present invention.

[0033] Figure 9 This is a flowchart illustrating the autonomous obstacle avoidance method provided in an embodiment of the present invention.

[0034] Figure 10 This is a flowchart illustrating the personnel perception safety protection method provided in an embodiment of the present invention.

[0035] Figure 11 This is a flowchart illustrating the self-positioning and support method for the stent provided in an embodiment of the present invention.

[0036] Figure 12 This is a schematic diagram of the control system of a hydraulic support provided in an embodiment of the present invention.

[0037] Figure 13 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0038] Detailed reference is made to exemplary embodiments, examples of which are illustrated in the accompanying drawings. The following description refers to the accompanying drawings, wherein the same reference numerals in different drawings denote the same or similar elements unless otherwise stated. In the following description of exemplary embodiments, the described embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with several aspects of the invention as described in the appended claims.

[0039] The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. In this invention and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein is intended to represent and include any or all possible combinations of one or more of the associated listed items.

[0040] It should be understood that although the terms “first,” “second,” “third,” etc., may be used herein to describe various types of information, such information should not be limited by these terms. These terms are only used to distinguish the types of information. For example, without departing from the scope of the invention, first information may be referred to as second information, and similarly, second information may be referred to as first information. The term “if” as used herein may be understood, depending on the context, to mean “at the time,” “at this moment,” or “in response to a judgment.”

[0041] Figure 1 This is a flowchart illustrating a control method for a hydraulic support according to an embodiment of the present invention. The hydraulic support includes a traveling mechanism, a top beam, a hydraulic column, and a sensing system. The sensing system includes a pose detection module and an environmental sensing module. Figure 1 As shown, the method includes the following steps: S1. Obtain the tunnel environment information collected by the environment perception module, extract tunnel features based on the tunnel environment information, and initialize and calibrate the initial attitude data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support, and use the initial pose information as the starting reference for the pose detection module to perform dead reckoning. S2. Based on the initial pose information, control the hydraulic column to release the support of the roof, and drive the traveling mechanism to move the hydraulic support along the roadway direction; S3. During the movement, the environmental perception module collects tunnel environmental perception data in real time and extracts tunnel environmental features based on the tunnel environmental perception data. S4. The pose detection module performs dead reckoning starting from the initial reference to obtain the dead reckoning position; the cumulative error of the dead reckoning position is corrected using the tunnel environment features to obtain the corrected positioning coordinates. S5. Control the hydraulic support to stop according to the corrected positioning coordinates, and control the hydraulic column to rise after stopping to complete the support of the top plate.

[0042] The hydraulic support control method provided in this invention utilizes roadway environmental characteristics to initialize and calibrate the pose detection module, providing a precise starting reference for the inertial navigation system (INS) to solve the initial alignment problem in the absence of GPS underground. During the support relocation process, the roadway environment is sensed in real time and features are extracted to correct the cumulative drift caused by the INS' long-distance operation, ensuring that the positioning accuracy continuously meets the support spacing requirements. Finally, automatic stopping and support are performed based on the corrected coordinates. Each step is tightly coupled through intermediate variables such as the "starting reference," "roadway environmental characteristics," and "corrected positioning coordinates," achieving physical isolation between operators and hazardous areas and fully automated support relocation throughout the entire process. This effectively solves the core problems of existing technologies, such as high personnel safety risks, reliance on manual experience for positioning, and the inability to self-correct INS drift, providing a safe and efficient intelligent control solution for advanced support in coal mine roadways.

[0043] In some embodiments, the tunnel environment information includes any one or more of the following: a preset tunnel marker image, a tunnel sidewall outline, a tunnel cross-sectional shape, or fixed equipment features; the tunnel features include any one or more of the following: visual features, geometric contour features, or radar reflection features.

[0044] Specifically, the different types of roadway environmental information mentioned above can be obtained by different sensors in the environmental perception module (such as cameras collecting identification images and radar collecting sidewall contours). The different types of roadway features can complement and verify each other, providing diverse references for subsequent initial calibration, thereby improving the robustness and environmental adaptability of the initial positioning, and ensuring that the initial pose can still be accurately determined under complex working conditions such as missing roadway markings and dust obstruction.

[0045] like Figure 2 As shown, in some embodiments, the initialization calibration of the initial pose data detected by the pose detection module based on the roadway features to determine the initial pose information of the hydraulic support includes: S11. Determine the absolute coordinates of at least one reference target among the tunnel features in the tunnel coordinate system; S12. Obtain the initial posture data detected by the pose detection module; S13. Based on the absolute coordinates of the reference target and the initial attitude data detected by the pose detection module, calculate the initial position coordinates and initial heading angle of the hydraulic support in the roadway coordinate system, and use the initial position coordinates and initial heading angle as the initial pose information.

[0046] Specifically, by using a reference target with known absolute coordinates in the tunnel environment (such as a preset marker or fixed equipment) as an external absolute reference, the relative attitude data output by the pose detection module is associated with the absolute position in the actual physical space. This anchors the origin of the pose detection module's coordinate system to the tunnel coordinate system, solving the problem that the initial value of the inertial navigation system cannot be obtained on its own, and providing an accurate starting reference for subsequent dead reckoning.

[0047] In some embodiments, the environmental perception module includes a ranging sensor, a camera, and an infrared detection device.

[0048] Specifically, through the coordinated configuration of multiple sensing modules, a comprehensive and multi-dimensional perception of the surrounding environment of the hydraulic support is achieved. The ranging sensor is used to perceive the spatial position and distance of obstacles, the camera is used to acquire visual details of the tunnel environment, and the infrared detection device is used to perceive the presence of personnel in the danger zone. The three types of sensing data are collected synchronously in time and complement each other in space, providing multi-dimensional data support for subsequent positioning correction, obstacle avoidance control and personnel safety protection.

[0049] In some embodiments, the ranging sensor is a millimeter-wave radar, which is deployed in at least four directions: front, rear, left, and right of the hydraulic support; the camera is deployed in at least four directions: front, rear, left, and right of the hydraulic support; and the infrared detection device is deployed in at least the front and rear directions of the hydraulic support.

[0050] Specifically, the millimeter-wave radar possesses anti-dust, anti-darkness, and anti-interference capabilities, enabling it to stably detect obstacle distances in the harsh environment of underground coal mines. The four-directional deployment of cameras and millimeter-wave radar ensures no blind spots in front, behind, or to the sides during support relocation. The infrared detection device identifies personnel based on human body temperature radiation characteristics, unaffected by dim lighting or dust obstruction underground. Its front-to-back deployment targets the most likely dangerous areas for personnel to enter during support relocation. This deployment scheme is fully adapted to the complex working conditions of narrow, dusty, and dimly lit coal mine roadways, providing a reliable hardware foundation for all-weather environmental perception and safety protection.

[0051] like Figure 3 As shown, in some embodiments, the method further includes: S6. During the movement, safety judgments are made in real time based on the roadway environment perception data collected by the environment perception module. S7. Based on the safety assessment results, control the hydraulic support to execute the corresponding obstacle avoidance strategy to ensure the safety of the support operation.

[0052] Specifically, during the relocation process, safety control strategies are implemented to ensure that there are no personnel or obstacles in the danger zone and that the equipment is in a normal posture when the support is moved. This avoids rash actions when there are potential safety hazards and ensures the safety of the relocation operation.

[0053] like Figure 4 As shown, in some embodiments, the safety determination based on the roadway environmental perception data collected by the environmental perception module includes: S61. Visual images are simultaneously acquired by the camera installed on the hydraulic support, infrared thermal images are simultaneously acquired by the infrared detection device, and the surrounding environment is continuously scanned by the millimeter-wave radar to obtain millimeter-wave radar ranging data. In step S61, the camera, infrared detection device, and millimeter-wave radar can simultaneously collect data, ensuring that the three sensing modules acquire multi-dimensional information about the tunnel environment at the same moment. This simultaneous data acquisition provides a spatiotemporal alignment basis for subsequent multi-sensor data fusion, avoiding data inconsistencies caused by time deviations.

[0054] S62. Determine obstacle features based on the visual image, and correct the obstacle features based on the millimeter-wave radar ranging data to obtain the true physical features of the obstacle; In step S62, in the dim, dusty environment of an underground coal mine, monocular or binocular vision has inherent biases in depth estimation and scale restoration. This step corrects the obstacle features extracted from the visual image by using high-precision ranging data from millimeter-wave radar (accuracy error not exceeding 10cm) as the true distance value, thus restoring the true size, shape features, and three-dimensional spatial distribution of the obstacles.

[0055] S63. Determine whether personnel have entered the dangerous area based on the infrared thermal imaging and the visual image. In step S63, the infrared detection device has the ability to identify human body temperature and can accurately capture human infrared radiation signals, unaffected by environmental factors such as dim lighting and dust in the mine. Through pixel-by-pixel binarization segmentation of the thermal imaging matrix (based on the human body surface temperature range) and connected component morphology discrimination, initial screening and separation of miscellaneous heat sources and human heat sources in the mine can be achieved. Relying on a 360° panoramic surround-view vision system formed by multiple cameras, machine vision algorithms are used to perform secondary verification of suspected targets screened by the infrared end. Visual target detection (using a lightweight YOLOv5s algorithm, targeting features such as safety helmets, tool outlines, and human limb structures) further confirms the initial infrared screening results, effectively eliminating infrared interference signals from underground equipment, high-temperature pipelines, etc.

[0056] S64. Make a safety determination based on the actual physical characteristics of the obstacle and the result of the personnel intrusion determination in the dangerous area.

[0057] Specifically, when dangerous conditions such as unauthorized personnel intrusion or approaching obstacles are detected, emergency stop, deceleration, and alarm actions will be triggered before manual remote control commands. This step integrates the actual physical characteristics of the obstacle with the personnel intrusion detection results to form a dual safety judgment logic of "obstacle avoidance + personnel protection"—if any condition triggers a hazard (obstacle too close or personnel entering a dangerous area), the system will execute the corresponding safety action. Compared to single-dimensional safety control (obstacle avoidance only or personnel protection only), the comprehensive judgment achieves a full assessment of the safety status of the support's operating environment.

[0058] like Figure 5 As shown, in some embodiments, determining obstacle features based on the visual image includes: S621. Based on the camera intrinsic parameters of each camera, lens distortion correction is performed on the visual image captured by the camera to obtain the corrected visual image. In step S621, to eliminate the imaging distortion caused by lens distortion and viewing angle misalignment, lens distortion correction is performed based on camera intrinsic parameters. The correction formula is as follows:

[0059] In the formula: These are the pixel coordinates of the distorted image (the original visual image acquired). For the corrected pixel coordinates, The radial distortion coefficient is... The tangential distortion coefficient is... This represents the distance from the pixel to the optical center.

[0060] S622. Feature points of the overlapping region of the corrected visual images corresponding to adjacent cameras are extracted using a feature matching algorithm. The homography matrix is ​​solved based on the feature points. The corrected visual images are then registered based on the homography matrix to obtain the image to be stitched. In step S622, after completing the multi-channel image distortion correction, feature points in the overlapping areas of adjacent camera images are extracted using a feature matching algorithm. The homography matrix is ​​then calculated to complete multi-view image registration, achieving spatial alignment of the four local images. The homography transformation formula is:

[0061] Normalization:

[0062] In the formula, These are the coordinates of the distortion-free pixels after distortion correction for a single-channel image. These are the pixel coordinates for viewpoint registration after homography transformation. These are the normalized pixel coordinates. It is a 3×3 homography matrix used to characterize the spatial mapping relationship between the viewpoints of adjacent cameras, so as to achieve accurate alignment of multiple images.

[0063] S623. The pixel-weighted fusion algorithm is used to stitch the image to be stitched to eliminate stitching gaps and brightness abrupt changes, and generate a 360° panoramic view image around the hydraulic support. In step S623, a pixel-weighted fusion algorithm is used to smooth the transition of overlapping areas in the image, eliminating stitching gaps and abrupt changes in brightness. The weighted fusion formula is as follows:

[0064] In the formula, The pixel grayscale values ​​of the merged panoramic image. , These are the pixel values ​​of the overlapping area in the images after registration of two adjacent paths. For adaptive pixel fusion weights, they are dynamically allocated based on the location of overlapping areas to eliminate splicing gaps and brightness abrupt changes.

[0065] Through the multi-camera stitching algorithm described above, the system ultimately generates a complete and seamless 360° panoramic view image around the bracket, effectively eliminating blind spots in visual detection.

[0066] S624. Use an edge detection algorithm to extract obstacle features from the 360° panoramic view image.

[0067] Specifically, the Canny edge detection algorithm can be used to extract the appearance features such as the edge contour, size and shape of obstacles to complete the obstacle shape reconstruction.

[0068] In some embodiments, the step of correcting the obstacle features based on the millimeter-wave radar ranging data to obtain the true physical features of the obstacle includes: transforming the obstacle features from the pixel coordinate system to the underground physical coordinate system based on the millimeter-wave radar ranging data and the pixel spatial coordinate mapping algorithm, and based on the camera's internal and external parameter calibration parameters, to obtain the true physical features of the obstacle.

[0069] Specifically, millimeter-wave radars in four directions (front, back, left, and right) can continuously scan the surrounding environment, collecting real-time distance and relative motion data to obstacles such as tunnel walls, waste rock, and temporary equipment, providing a high-precision distance benchmark for spatial location determination. Further, the high-precision ranging data from the millimeter-wave radar is deeply integrated with a pixel spatial coordinate mapping algorithm. Based on the camera's internal and external parameter calibration parameters, the conversion between the pixel coordinate system and the underground physical coordinate system is completed. The algorithm mapping relationship is as follows:

[0070] In the formula: Image pixel coordinates, The three-dimensional physical coordinates of the obstacle in the well. The pixel coefficient is the camera focal length. The principal pixel coordinates These are the rotation and translation matrices of the camera relative to the support body, respectively. This refers to the camera's depth-of-field scale parameters. The coordinate transformation relationship is solved based on the camera's intrinsic and extrinsic calibration parameters, while the measured distance from the millimeter-wave radar is used as... The system uses true axis constraint values ​​to correct depth-of-field deviations and scale errors caused by pure visual positioning, restoring the true size, shape characteristics, and three-dimensional spatial distribution of obstacles. This assists the controller in determining the type of obstacle (movable / fixed) and provides a basis for obstacle avoidance decisions. During this process, the inertial navigation system updates the support posture data in real time, dynamically correcting detection errors and ensuring the reliability of perception data across the entire scene.

[0071] like Figure 6 As shown, in some embodiments, the step of determining whether personnel have entered a dangerous area based on the infrared thermal image and the visual image includes: S631. Filter the pixels in the infrared thermal image based on the temperature threshold to obtain the suspected human targets in the infrared thermal image. In step S631, the infrared detection device continuously scans the preset danger zones in front of and behind the support, continuously collects infrared radiation light signals from various objects within the zone, and converts the light signals into resolvable electrical signals, ultimately generating a thermal imaging matrix with pixel-by-pixel temperature parameters. Based on the stable temperature characteristics of the human body surface, a dedicated human body temperature discrimination range is constructed, and pixel-by-pixel binarization segmentation processing is performed on the infrared imaging matrix to achieve initial screening and separation of human body heat sources from miscellaneous heat sources underground. By limiting the effective human body temperature range, invalid interfering heat sources such as low-temperature tunnel walls, normal-temperature mechanical equipment, and high-temperature pipelines are filtered out, completing the preliminary extraction of suspected human target areas.

[0072] In the formula: The result is the pixel binarization determination. Coordinates in the thermal imaging matrix The measured temperature corresponding to the pixel and These are the upper and lower limits of the effective human body temperature. This indicates that the pixel is a suspected human heat source. This indicates that the pixel is an invalid source of interference heat.

[0073] S632. The suspected human targets are screened using a connected component morphological feature discrimination mechanism to obtain human target screening results; In step S632, the image after temperature threshold segmentation still contains point-like high-temperature noise and scattered small-area heat source false targets, which can easily cause false detections by the system. Therefore, a connected component morphological feature discrimination mechanism can be adopted. By statistically analyzing the pixel connectivity area of ​​the effective temperature region and combining it with the contour scale features of the downhole worker, a minimum effective area threshold is set to eliminate isolated noise and invalid small-area false targets, thus completing a secondary, precise screening of human targets.

[0074] In the formula: The result of the target shape determination. This represents the total area of ​​connected pixels after temperature filtering. The minimum effective contour area threshold for underground personnel. For effective human target, This is an invalid spurious target.

[0075] S633. Identify human targets within the danger zone of the 360° panoramic view image using machine vision algorithms to obtain human target identification results; In step S633, relying on a 360° panoramic surround-view vision system formed by stitching together multiple cameras, machine vision algorithms are used to perform secondary verification on suspected targets screened by the infrared end, achieving accurate identification of personnel targets. Addressing issues such as darkness, dust obstruction, and blurred images underground, the real-time video stream captured by the cameras is first preprocessed. Noise reduction filtering, haze removal, and dynamic brightness adaptive adjustment algorithms are used to optimize image quality and restore effective scene feature information. After preprocessing, the system uses a lightweight YOLOv5s target detection algorithm as the core recognition model. It extracts features and identifies targets based on the specific characteristics of underground workers (safety helmet, workwear outline, human limb structure), accurately distinguishing workers from common underground interfering objects such as supports, anchor bolts, conveyors, and coal blocks. Simultaneously, the system delineates an effective detection area around the supports, identifying and judging targets only within this dangerous area, filtering out invalid personnel targets outside the area to avoid false triggering. Finally, based on the target feature matching degree, the system outputs the confidence level of personnel presence at the visual end, providing a basis for data fusion and judgment.

[0076] S634. Based on the human target screening results and human target recognition results, determine the intrusion of personnel into the dangerous area.

[0077] In step S634, a weighted fusion algorithm can be used to fuse the dual-source sensing data, avoiding the identification defects of a single detection method, eliminating infrared signal interference from underground equipment, debris, etc., and confirming the authenticity of the personnel target. The fusion algorithm formula is as follows:

[0078] in, The probability of comprehensive judgment of personnel targets; Weights for infrared detection data; To determine the probability of a person's presence using an infrared detection device; The weights for the camera image data are given, and satisfy the following conditions: ; To determine the probability of a person's presence based on image recognition, when At that time, it was confirmed that there were personnel present.

[0079] like Figure 7 As shown, in some embodiments, the step of performing dead reckoning using the pose detection module with the starting reference as the starting point to obtain the dead reckoning position includes: S41. Take the initial position coordinates in the initial pose information as the integration starting point, and take the initial heading angle in the initial pose information as the initial heading. S42. The posture detection module collects the acceleration data of the hydraulic support along the roadway direction in real time. S43. Perform a second integration on the acceleration data to obtain the displacement increment of the hydraulic support relative to the initial reference, and accumulate the displacement increment to the initial position coordinates to obtain the real-time calculated position coordinates.

[0080] Specifically, starting from a known initial state, the process of continuously calculating real-time position and heading using the basic principles of inertial navigation (acceleration integration to obtain velocity, and re-integration to obtain displacement) has the advantages of high-frequency output, good continuity, and immunity to external environmental interference, providing a foundation of position data to be corrected for subsequent error correction using tunnel environmental characteristics.

[0081] like Figure 8 As shown, in some embodiments, the step of correcting the cumulative error of the dead reckoning position using the tunnel environment characteristics to obtain the corrected positioning coordinates includes: S461. Based on the geometric contour features and / or spatial distribution features in the roadway environment, and combined with the distance data between the hydraulic support and the roadway sidewall collected in real time by the ranging sensor, calculate the first absolute position of the hydraulic support in the transverse direction of the roadway. S462. Based on the visual features in the tunnel environment, identify the preset support mileage scale or spacing mark in the tunnel through the environmental image collected by the camera, and obtain the second absolute position of the hydraulic support in the longitudinal direction of the tunnel. S463. The lateral coordinates of the dead reckoning position are weighted and fused with the first absolute position to obtain the corrected lateral positioning coordinates; S464. The longitudinal coordinates of the dead reckoning position are weighted and fused with the second absolute position to obtain the corrected longitudinal positioning coordinates; S465. Combine the corrected lateral positioning coordinates and the corrected longitudinal positioning coordinates to form the corrected positioning coordinates.

[0082] Specifically, the first absolute position is calculated laterally by combining geometric contour features with distance measured by the ranging sensor, and then weighted and fused with the dead reckoning lateral coordinates. The second absolute position is obtained longitudinally by identifying mileage scales or spacing markers through visual features, and then weighted and fused with the dead reckoning longitudinal coordinates. This independent lateral and longitudinal fusion correction mechanism targets different types of drift errors generated by the inertial navigation system in different directions, effectively solving the problem of excessive heading deviation and mileage accumulation deviation after long-distance operation using pure inertial navigation.

[0083] To better understand the present invention, the control method of the hydraulic support provided by the present invention will be described in detail below through a specific embodiment.

[0084] The existing advanced support operation has the following technical problems: (1) It relies on operators to enter the dangerous area for a long time to carry out on-site operations, facing safety hazards such as roof collapse, collision and squeezing; (2) It lacks effective environmental perception and attitude monitoring methods. During the movement of the frame, it is impossible to identify the surrounding obstacles, personnel status and its own attitude in real time, which is prone to collision and misoperation; (3) It mainly relies on manual experience for positioning and operation, which is slow and difficult to achieve accurate positioning and rapid support. Moreover, repeated application of concentrated loads is prone to causing disturbance and damage to the roadway anchor bolt and cable support system; (4) It is difficult to achieve remote and automated control, which cannot meet the needs of unmanned operation in dangerous areas.

[0085] To address the aforementioned issues, this embodiment provides an intelligent tracked self-propelled hydraulic support solution: the entire machine relies on a tracked walking structure to adapt to complex working conditions such as uneven underground floors and narrow, variable tunnels, achieving roof support and load-bearing through hydraulic columns and top beams; the support is equipped with a multi-sensor fusion perception system, integrating omnidirectional millimeter-wave radar (obstacle ranging), 360° surround-view industrial cameras (environmental panoramic imaging), front and rear infrared personnel detection devices (personnel intrusion detection in dangerous areas), and a central inertial navigation system (support attitude and heading monitoring), providing comprehensive data support for intelligent safety control. The system uses an on-board controller as the core control unit, integrating wireless remote control commands and real-time perception data from multiple sensors, and relying on multi-level intelligent algorithms to complete environmental perception, risk assessment, attitude correction, and closed-loop control of support position. During operation, the system can adaptively perform safety actions such as emergency stop protection, deceleration and obstacle avoidance, and attitude correction based on infrared personnel detection, radar obstacle detection, and inertial navigation attitude detection results. It also supports operators to complete the operation in a safe area via wireless remote control or automatic control, realizing an automated operation process of "removing support - tracked frame shifting - precise positioning - re-supporting". Furthermore, it can rely on autonomous control algorithms to achieve unmanned intelligent support, thereby reducing the frequency of personnel entering dangerous areas, reducing the risk of roof instability and safety accidents, improving support efficiency, and reducing repeated disturbances to the roadway anchor bolt and cable structure. It provides a safe, accurate, efficient, and intelligent support solution for coal mine roadway pre-support, roadway support, and complex underground space support scenarios.

[0086] Based on the overall hardware architecture and control logic described above, the following sections will detail the specific implementation scheme and technical principles of this invention from the perspectives of three core functions: autonomous obstacle avoidance for the support structure, personnel safety protection, and precise positioning and support.

[0087] I. Autonomous obstacle avoidance function

[0088] The autonomous obstacle avoidance function is used to address issues such as uneven roadway floors, protruding roadway sidewalls, and obstacles like gangue and temporary equipment in coal mine roadway pre-support and roadway support scenarios. It ensures that the support frame does not collide during pre-support relocation and roadway support adjustment, guaranteeing equipment safety and the continuity of support operations. (See also...) Figure 9 The specific implementation method is as follows: (1) Hardware configuration Using millimeter-wave radar and industrial cameras as obstacle avoidance sensing components, combined with an inertial navigation system, a multi-dimensional obstacle avoidance sensing system is constructed. This system is suitable for complex environments such as narrow spaces, dust, and dim lighting in coal mine roadway pre-support and roadway support, and meets the obstacle avoidance requirements during the relocation process.

[0089] Millimeter-wave radar: Installed in the front, back, left, and right directions of the support, it has the ability to resist dust, darkness, and interference. It is suitable for harsh environments such as advanced support and roadway support. It can detect the distance and relative speed between the support and obstacles such as roadway walls, gangue, and temporary support equipment in real time. The detection range covers 0.5-8 meters around the support (suitable for narrow spaces in roadways). The detection accuracy error does not exceed 10cm, ensuring the accuracy of obstacle avoidance. Industrial cameras: Working in conjunction with millimeter-wave radar, they collect 360° panoramic images of the bracket's surroundings, presenting the specific shape and location of obstacles in real time, assisting the controller in determining the type of obstacle (movable / fixed), and providing a basis for obstacle avoidance decisions; Inertial navigation system: assists in detecting the walking posture, tilt angle and heading information of the support, avoids the deviation of obstacle avoidance judgment caused by the tilt of the support, and ensures the accuracy of obstacle avoidance actions.

[0090] (2) Working principle and process

[0091] The autonomous obstacle avoidance function is implemented through a process of "obstacle detection—distance judgment—obstacle avoidance decision—action execution," combined with preset safety logic, to achieve intelligent obstacle avoidance in different scenarios. Details are as follows: During the pre-support relocation and track support adjustment of the scaffold, a 360° panoramic surround-view detection system is constructed using a four-camera full-area deployment and image stitching fusion algorithm. Four explosion-proof industrial cameras (front, rear, left, and right) simultaneously acquire on-site images, achieving blind-spot-free environmental perception. To eliminate imaging distortion caused by lens distortion and viewing angle misalignment, the system first performs lens distortion correction based on camera intrinsic parameters. The correction formula is as follows:

[0092] In the formula: These are the pixel coordinates of the distorted image. For the corrected pixel coordinates, The radial distortion coefficient is... The tangential distortion coefficient is... This represents the distance from the pixel to the optical center.

[0093] After completing multi-channel image distortion correction, the system extracts feature points in the overlapping areas of adjacent camera images using a feature matching algorithm, solves the homography matrix to complete multi-view image registration, and achieves spatial alignment of four local images. The homography transformation formula is as follows:

[0094] Normalization:

[0095] In the formula, These are the coordinates of the distortion-free pixels after distortion correction for a single-channel image. These are the pixel coordinates after homography transformation to complete viewpoint registration. These are the normalized pixel coordinates. It is a 3×3 homography matrix used to characterize the spatial mapping relationship between the viewpoints of adjacent cameras, so as to achieve accurate alignment of multiple images.

[0096] Finally, a pixel-weighted fusion algorithm is used to smooth the transition of overlapping areas in the image, eliminating stitching gaps and abrupt brightness changes. The weighted fusion formula is as follows:

[0097] In the formula, The pixel grayscale values ​​of the merged panoramic image. , These are the pixel values ​​of the overlapping area in the images after registration of two adjacent paths. For adaptive pixel fusion weights, they are dynamically allocated based on the location of overlapping areas to eliminate splicing gaps and brightness abrupt changes.

[0098] Using the multi-camera stitching algorithm described above, the system ultimately generates a complete and seamless 360° panoramic view image of the bracket's perimeter, effectively eliminating blind spots in visual detection. Based on this, the Canny edge detection algorithm is used to extract appearance features such as obstacle edge contours, dimensions, and shapes, completing the obstacle shape reconstruction.

[0099] The system continuously scans the surrounding environment using millimeter-wave radar in four directions (front, back, left, and right), collecting real-time distance and relative motion data to obstacles such as tunnel walls, waste rock, and temporary equipment, providing a high-precision distance reference for spatial location determination. The system further integrates high-precision ranging data from the millimeter-wave radar with a pixel spatial coordinate mapping algorithm, using camera intrinsic and extrinsic calibration parameters to complete the conversion between the pixel coordinate system and the underground physical coordinate system. The algorithm mapping relationship is as follows:

[0100] In the formula: Image pixel coordinates, The three-dimensional physical coordinates of the obstacle in the well. The pixel coefficient is the camera focal length. The principal pixel coordinates These are the rotation and translation matrices of the camera relative to the support body, respectively. This refers to the camera's depth-of-field scale parameters. The system solves for the coordinate transformation relationship based on the camera's intrinsic and extrinsic calibration parameters, while also using the measured distance from the millimeter-wave radar as a reference. The system uses true axis constraint values ​​to correct depth-of-field deviations and scale errors caused by pure visual positioning, restoring the true size, shape characteristics, and three-dimensional spatial distribution of obstacles. This assists the controller in determining the type of obstacle (movable / fixed) and provides a basis for obstacle avoidance decisions. During this process, the inertial navigation system updates the support posture data in real time, dynamically correcting detection errors and ensuring the reliability of perception data across the entire scene.

[0101] Distance threshold judgment: Considering the narrow space characteristics of coal mine roadway pre-support and haulage roadway support, the controller presets two obstacle avoidance safety thresholds (early warning threshold and emergency threshold), where the early warning threshold is 1.0~1.5 meters and the emergency threshold is 0.5~0.8 meters; a distance deviation early warning algorithm is used to judge the obstacle risk level in real time. The algorithm formula is as follows:

[0102] in: The obstacle risk coefficient; This is the safety threshold (early warning threshold or emergency threshold) for the current scenario. This represents the actual distance to the obstacle detected by the millimeter-wave radar. When the obstacle distance is higher than the safety threshold, the support walks normally; when When, trigger an alert; when At that time, emergency obstacle avoidance is triggered.

[0103] The system makes intelligent obstacle avoidance decisions based on the risk coefficient output by the algorithm and the type of obstacle. When the early warning mechanism is triggered, the system activates an audible and visual warning to alert on-site personnel to surrounding obstacles, while simultaneously controlling the support frame to reduce its movement speed for stable operation at low speed, avoiding potential collision risks. When the emergency obstacle avoidance mechanism is triggered, the controller immediately locks the support frame's movement and support actions, forcing the equipment to stop and preventing collisions. For detected fixed obstacles, the system plans the optimal detour and fine-tunes the attitude path; for movable obstacles, it prioritizes warning personnel to clear them, achieving precise obstacle avoidance control adapted to different scenarios. Once the obstacle is removed, the detection distance returns to a safe range, and the risk coefficient falls back to a safe level, the system automatically removes the restrictions, and the support frame resumes normal operation, ensuring continuous and stable operation of the support work.

[0104] II. Implementation of Personnel Perception Safety Protection Function

[0105] Personnel detection is the core guarantee for the safe operation of supports in coal mine roadway pre-support and haulage support scenarios. The core objective is to detect in real time whether personnel have entered the hazardous area of ​​the support, promptly triggering safety protection mechanisms to avoid personnel casualties. This adapts to the complex environment of underground support operations. (See also...) Figure 10 The specific implementation method is as follows: (1) Hardware configuration An infrared detection device is used as the core sensing component, combined with the synergistic effect of a multi-sensor sensing system, to build a comprehensive personnel detection network.

[0106] Infrared detection device: Installed at key positions in the front and rear directions of the support, it adopts a high-precision infrared thermal imaging sensor, has the ability to identify human body temperature, and can accurately capture human infrared radiation signals, unaffected by environmental factors such as dim lighting and dust in the well. Collaborative sensing components: Industrial cameras around the bracket assist in collecting personnel image information, complementing infrared detection data, improving the accuracy of personnel detection, and avoiding misjudgments caused by foreign object interference; (2) Working principle and process The infrared personnel detection device continuously scans the pre-defined danger zones before and after the support (in the context of coal mine roadway pre-support and haulage support scenarios, the danger zone is pre-defined as 1.5-2.5 meters around the support, suitable for the narrow space of the haulage and the activity range of personnel working on pre-support), continuously collecting infrared radiation light signals from objects within the zone, converting the light signals into resolvable electrical signals, and finally generating a thermal imaging matrix with pixel-by-pixel temperature parameters. Based on the stable characteristics of human body surface temperature, a dedicated human body temperature discrimination range is constructed, and pixel-by-pixel binarization segmentation processing is performed on the infrared imaging matrix to achieve initial screening and separation of human body heat sources from miscellaneous heat sources underground. By limiting the effective human body temperature range, invalid interfering heat sources such as low-temperature roadway walls, normal-temperature mechanical equipment, and high-temperature pipelines are filtered out, completing the preliminary extraction of suspected human target areas.

[0107] In the formula: The result is the pixel binarization determination. Coordinates in the thermal imaging matrix The measured temperature corresponding to the pixel and These are the upper and lower limits of the effective human body temperature. This indicates that the pixel is a suspected human heat source. This indicates that the pixel is an invalid source of interference heat.

[0108] Even after temperature threshold segmentation, images still contain point-like high-temperature noise and scattered small-area heat source false targets, which can easily lead to false detections by the system. Therefore, this system employs a connected component morphological feature discrimination mechanism. By statistically analyzing the pixel connectivity area of ​​the effective temperature region and combining it with the contour scale characteristics of downhole workers, a minimum effective area threshold is set to eliminate isolated noise and invalid small-area false targets, thus achieving a second, more precise screening of human targets.

[0109] In the formula: The result of the target shape determination. This represents the total area of ​​connected pixels after temperature filtering. The minimum effective contour area threshold for underground personnel. For effective human target, This is an invalid spurious target.

[0110] Relying on a 360° panoramic surround-view vision system formed by stitching together multiple cameras, the system uses machine vision algorithms to perform secondary verification on suspected targets screened by the infrared end, achieving accurate identification of personnel targets. Addressing issues such as darkness, dust obstruction, and blurred images underground, the system first preprocesses the real-time video streams captured by the cameras. Noise reduction filtering, haze removal, and dynamic brightness adaptive adjustment algorithms optimize image quality, restoring effective scene feature information. After preprocessing, the system uses a lightweight YOLOv5s target detection algorithm as the core recognition model. It extracts features and identifies targets based on the specific characteristics of underground workers (safety helmet, workwear outline, human limb structure), accurately distinguishing workers from common underground interfering objects such as supports, anchor bolts, conveyors, and coal blocks. Simultaneously, the system delineates an effective detection area around the supports, identifying and judging targets only within this hazardous area, filtering out invalid personnel targets outside the area to avoid false triggering. Finally, based on the degree of target feature matching, the system outputs the confidence level of personnel presence at the visual end, providing a basis for data fusion and judgment.

[0111] The system employs a weighted fusion algorithm to fuse dual-source sensing data, avoiding the identification defects of single detection methods, eliminating infrared signal interference from underground equipment, debris, and other sources, and confirming the authenticity of personnel targets. The fusion algorithm formula is as follows:

[0112] in, The probability of comprehensive judgment of personnel targets; Weights for infrared detection data; To determine the probability of a person's presence using an infrared detection device; The weights for the camera image data are given, and satisfy the following conditions: ; To determine the probability of a person's presence based on image recognition, when At that time, it was confirmed that there were personnel present.

[0113] The support controller determines whether personnel have entered the danger zone based on preset safety logic. If personnel are confirmed to have entered the danger zone, a safety control command is immediately triggered. The system immediately performs an emergency stop operation, stopping all movements of the tracked walking mechanism and the extension and retraction of the hydraulic column to prevent personnel from being crushed or collided during support movement or support actions. An audible and visual alarm signal is also issued to remind operators and on-site personnel to evacuate the danger zone. The emergency stop state can only be lifted and the support can resume normal operation after personnel have left the danger zone.

[0114] III. Self-positioning and support function of the scaffold

[0115] The precise positioning and support function ensures that the supports can be accurately moved to the preset support positions for pre-support and roadway support in coal mine roadways. This guarantees the positioning accuracy of the pre-support and the spacing accuracy of the roadway support, avoiding support failure and roadway roof collapse risks caused by positioning deviations. Considering the environmental characteristics of pre-support and roadway support in underground coal mines, a positioning method combining inertial navigation and multi-sensor fusion is adopted. (See [link to documentation]). Figure 11 The specific implementation is as follows: (1) Hardware configuration Using an inertial navigation system as the core positioning component, combined with millimeter-wave radar, industrial cameras, and pre-set positioning benchmarks for advanced support and roadway support, a high-precision positioning system is constructed to solve the problem of missing GPS signals downhole and to meet the precise positioning requirements of advanced support and roadway support.

[0116] Inertial navigation system: Installed in the central area of ​​the support body, it has attitude detection, tilt angle detection, heading detection and position calculation functions. It can collect the acceleration and angular velocity data of the support in real time, and obtain the real-time position, heading and tilt angle information of the support through integration calculation. It is not affected by underground dust and dim environment. Auxiliary positioning components: Millimeter-wave radar detects the distances to the sidewalls of the roadway, the fixed support equipment in the advanced support area, and the roadway outline to form a positioning reference; industrial cameras collect preset positioning markers (such as advanced support mileage markers and roadway support spacing markers) and roadway outline features in the roadway to help correct positioning deviations and ensure that the positioning fits the support scenario. Vehicle-mounted controller: It has a built-in positioning algorithm and preset position database, which stores the coordinate information of each predetermined support position of the pre-support and the roadway support. Combined with multi-sensor data, it realizes the calibration of positioning accuracy and position closed-loop control, and meets the continuous positioning requirements of pre-support and roadway support.

[0117] (2) Work process and principle

[0118] After the support is started, the inertial navigation system completes initialization. Combined with the preset positioning reference points (such as the positioning mark at the entrance of the roadway and the starting mark of the advance support) of the underground advance support and the roadway support, the initial position coordinates of the support are determined and uploaded to the vehicle controller as the positioning reference to ensure that the positioning starting point is consistent with the roadway support scenario.

[0119] During the movement of the support structure, the inertial navigation system continuously collects data on the support structure's acceleration, angular velocity, tilt angle, and heading. The inertial navigation positioning algorithm calculates the support structure's current position coordinates in real time. The core formula of the algorithm is as follows:

[0120] in, The coordinates of the current position of the support; These are the initial positioning coordinates; For stent in Real-time speed in direction; The initial velocity; Data collected for inertial navigation systems Directional acceleration in real time; t is the time of movement.

[0121] Pure inertial navigation relies on the time-series integration of acceleration and angular velocity to calculate position, offering advantages such as high-frequency output, good continuity, and immunity to external environmental interference. However, it suffers from accumulated drift errors over long periods of operation. In long-distance support operations in tunnels, inertial navigation positioning can exhibit problems such as heading deviation and excessive cumulative mileage deviation, leading to insufficient support advance distance and support spacing, thus failing to meet the requirements for long-term, high-precision support operations. Therefore, the system incorporates millimeter-wave radar and camera visual features for fusion calibration, providing real-time constraints and iterative corrections to the accumulated drift errors of the inertial navigation system.

[0122] Millimeter-wave radar is responsible for correcting lateral drift errors and for real-time monitoring of the relative distances between the support frame and the roadway sidewall, roadway profile, and fixed advance support equipment. Combined with the standard roadway width, the lateral distance between the support frame and the roadway wall is obtained by fitting the data. The camera vision system is responsible for compensating for cumulative longitudinal mileage errors. Through panoramic image feature matching, it identifies absolute benchmark labels such as preset support mileage scales and spacing markings in the tunnel, matches them with the support standard point database, and obtains the drift-free longitudinal absolute position. The system combines the current lateral and longitudinal absolute positions of the support obtained from millimeter-wave radar and camera vision recognition. It then employs a multi-sensor weighted fusion correction algorithm to iteratively compensate for the raw inertial navigation positioning data, achieving comprehensive error correction. The algorithm formula is as follows:

[0123] In the formula: The optimal positioning coordinates for the stent after fusion correction. Real-time position is calculated for inertial navigation dead reckoning. These are the weights for the horizontal and vertical inertial navigation data, respectively. For horizontal radar data weights, These are the weights for the vertical visual data. The weights for each dimension satisfy the normalization constraint: , The onboard controller uses this algorithm to independently weighted and fused the lateral and longitudinal positions, compensating in real time for the positioning drift error accumulated by the inertial navigation over the duration of operation, ensuring real-time positioning accuracy and meeting the needs of the pre-support operation in the roadway.

[0124] IV. Multi-sensor linkage and wireless remote control collaborative operation mechanism

[0125] During underground operations, the entire multi-sensor sensing system provides comprehensive environmental data and equipment status information for remote control, serving as a core prerequisite for safe remote control and intelligent operation. Millimeter-wave radar, industrial surround-view cameras, infrared personnel detection devices, and inertial navigation systems synchronously collect real-time information on roadway obstacles, personnel in hazardous areas, equipment attitude, and location around the clock. After being processed by the onboard controller's algorithm, this information accurately identifies on-site operational risks and equipment operating status. Operators remotely issue commands for moving, supporting, and starting / stopping equipment from a safe area underground via a wireless remote control system. The equipment no longer relies solely on manual operating experience but uses multi-source sensing data as a decision-making benchmark to assist humans in completing precise operations. Simultaneously, the sensing system monitors the working environment in real time throughout the process. Once it detects dangerous conditions such as personnel accidentally entering, obstacles approaching, or equipment attitude deviation, it will trigger emergency stop, deceleration, and alarm actions before human remote control commands, forming a safety redundancy backup for remote control operations.

[0126] Relying on the collaborative mechanism of perception and remote control, the equipment can realize the intelligent operation of the entire process of "remote control to remove support - real-time protection with perception - autonomous obstacle avoidance and frame relocation - precise positioning with perception - automatic locking of support", which greatly reduces the frequency of manual operation intervention. At the same time, it provides core perception data support for subsequent unmanned autonomous frame relocation and cluster collaborative support, truly realizing the safe, remote and intelligent upgrade of downhole support operations.

[0127] The hydraulic support control method proposed in this invention, through the integrated application of a tracked autonomous movement structure, a multi-sensor environmental perception system, a wireless remote control system, and intelligent safety control technology, has significant advantages in safety, intelligence, and engineering applications compared to traditional manual support movement and ordinary tracked hydraulic supports. Achieving sensory interlocking and remote control operation with human-machine separation significantly improves the inherent safety of support operations: Constructing a multi-dimensional integrated protection system that integrates infrared personnel detection, millimeter-wave radar obstacle perception, camera panoramic vision, and inertial navigation attitude monitoring, relying on multi-source sensing real-time interlocking logic combined with wireless remote control technology, it changes the traditional operation mode of close-range manual following of the support, allowing operators to stay away from the support operation area and complete the movement and support operations in a safe area, thereby reducing the frequency of personnel entering dangerous areas of the roof, areas of transfer equipment, and narrow tunnels, effectively reducing the risk of safety accidents such as collisions, crushing, and roof falling debris injuries.

[0128] A comprehensive perception and intelligent interlocking protection mechanism is constructed to solve the operational hazards caused by lack of perception and monitoring: Integrating infrared personnel detection, millimeter-wave radar obstacle detection, visual environmental perception and inertial navigation attitude monitoring functions, it can collect the attitude information of personnel, obstacles and equipment around the support in real time around the clock. Combined with intelligent interlocking control logic, it realizes risk warning, deceleration regulation and emergency protection, making up for the technical shortcomings of traditional equipment lack of perception, unmonitored status and unpredictable risks, and greatly improving the stability of equipment operation and operational safety.

[0129] Eliminating reliance on manual experience to achieve precise positioning and efficient support, adapting to the needs of long-distance roadway operations: Adopting inertial navigation-based precise positioning technology, integrating multiple sensors to correct displacement and attitude deviations in real time, it overcomes the drawbacks of traditional advanced support operations that heavily rely on manual experience for visual alignment and manual control. It enables rapid follow-up and relocation of supports, precise alignment and positioning, and timely pressure stabilization support during the advanced support process, effectively ensuring the standardization accuracy of the advanced support distance and support layout spacing in the roadway, significantly improving the efficiency of advanced support operations and the quality of support formation, and stably adapting to the engineering operation requirements of continuous cyclic advanced support in long-distance underground roadways.

[0130] Meanwhile, the embodiments of the present invention have strong intelligent expansion capabilities, and can be further combined with path planning, automatic navigation and cluster collaborative control technologies to realize the autonomous movement and unmanned operation of hydraulic supports, providing an important technical foundation for the intelligent, unmanned and unmanned construction of coal mines, and has good engineering promotion value and industrial application prospects.

[0131] Besides the tracked self-moving hydraulic support structure described in the above embodiments, those skilled in the art can also use other forms of moving mechanisms to achieve autonomous movement of the support without departing from the core ideas and objectives of this invention. For example, in other embodiments, wheeled drive structures, tracked and wheeled combination structures, sliding drive structures, track-guided moving structures, or hydraulic stepping moving structures can be used to move the support forward. The core purpose of all these is to achieve autonomous movement and re-support of the hydraulic support in the roadway. Therefore, the above structures are all equivalent alternatives to this invention.

[0132] In terms of environmental perception, this embodiment of the invention uses millimeter-wave radar, industrial cameras, infrared personnel detection devices, and inertial navigation systems to construct a multi-sensor fusion perception system. However, in other embodiments, lidar, ultrasonic sensors, 3D vision cameras, structured light sensors, thermal imaging devices, or UWB positioning systems can be used to replace some of the perception devices. As long as the detection, identification, or positioning of obstacles, personnel status, spatial environment, and support posture can be achieved, the core objective of this invention can be achieved, and therefore, they belong to the equivalent technical solutions of this invention.

[0133] In terms of control methods, the present invention adopts a combination of wireless remote control and automatic control. In other embodiments, wired remote control, ground centralized control, downhole centralized control system control, 5G communication control, or autonomous control based on artificial intelligence algorithms can also be used to realize the movement and support control of the support. The core still belongs to the category of remote, intelligent and unmanned control. Therefore, the above control methods are all technical extensions within the scope of protection of the present invention.

[0134] In terms of image processing and intelligent control, video recognition algorithms, digital twin systems, path planning algorithms, SLAM spatial mapping algorithms, or multi-device collaborative control algorithms can be further employed to optimize and control the support's operating path, obstacle avoidance, and support actions. Even if different software algorithms or data processing methods are used, as long as their core objectives remain improving the safety of hydraulic support movement, reducing personnel entry into dangerous areas, and achieving autonomous support, they should all be considered equivalent implementations of this invention.

[0135] Any equivalent replacement, functional combination, module addition or subtraction, or form change made to the support moving structure, environmental perception system, control system, and intelligent control algorithm based on the core technical ideas of this invention, as long as it can realize the autonomous movement, safe obstacle avoidance, remote control, and roof support functions of the hydraulic support, should fall within the protection scope of this invention.

[0136] Based on the same inventive concept, embodiments of the present invention also provide a control system for a hydraulic support.

[0137] Figure 12 This is a schematic diagram of the control system for a hydraulic support provided in an embodiment of the present invention. Figure 12 As shown in the figure, an embodiment of the present invention provides a control system for a hydraulic support, the hydraulic support including a traveling mechanism, a top beam and a hydraulic column, and the control system including: The sensing device 1 includes a pose detection module and an environmental sensing module. The environmental sensing module is used to collect roadway environmental information and to collect roadway environmental sensing data during the movement of the hydraulic support. The pose detection module is used to detect the attitude data of the hydraulic support and perform dead reckoning. Controller 2 is connected to the sensing device 1, the walking mechanism, and the hydraulic column via signals. Controller 2 is configured to perform the following steps: The system acquires the tunnel environment information collected by the environmental perception module, extracts tunnel features based on the tunnel environment information, and initializes and calibrates the initial attitude data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support. The initial pose information is then used as the starting reference for the pose detection module to perform dead reckoning. Based on the initial pose information, the hydraulic column is controlled to release the support of the roof, and the traveling mechanism is driven to move the hydraulic support along the roadway. During the movement, the environmental perception module collects tunnel environmental perception data in real time and extracts tunnel environmental features based on the tunnel environmental perception data. The pose detection module performs dead reckoning starting from the initial reference to obtain the dead reckoning position; the cumulative error of the dead reckoning position is corrected using the tunnel environment features to obtain the corrected positioning coordinates. The hydraulic support is stopped according to the corrected positioning coordinates, and the hydraulic column is raised after stopping to complete the support of the top plate.

[0138] The hydraulic support control system provided in this invention utilizes roadway environmental characteristics to initialize and calibrate the pose detection module, providing a precise starting reference for the inertial navigation system (INS) to solve the initial alignment problem in the absence of GPS underground. During the support relocation process, the system senses the roadway environment in real time and extracts features to correct the cumulative drift caused by the INS' long-distance operation, ensuring that the positioning accuracy continuously meets the support spacing requirements. Finally, the system automatically stops and supports the system based on the corrected coordinates. Each step is tightly coupled through intermediate variables such as the "starting reference," "roadway environmental characteristics," and "corrected positioning coordinates," achieving physical isolation between operators and hazardous areas and fully automated support relocation throughout the entire process. This effectively solves the core problems of existing technologies, such as high personnel safety risks, reliance on manual experience for positioning, and the inability to self-correct INS drift, providing a safe and efficient intelligent control solution for advanced support in coal mine roadways.

[0139] Figure 13 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, as shown below. Figure 13 As shown, the electronic device may include a processor 301, a communications interface 302, a memory 303, and a communication bus 304, wherein the processor 301, the communications interface 302, and the memory 303 communicate with each other via the communication bus 304. The processor 301 may call logical instructions in the memory 303 to execute the methods described in any of the above embodiments.

[0140] Furthermore, the logical instructions in the aforementioned memory 303 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0141] This embodiment of the invention provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments.

[0142] This embodiment provides a computer-readable storage medium storing a computer program that causes the computer to perform the methods provided in the above-described method embodiments.

[0143] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0145] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0146] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0147] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0148] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A control method for a hydraulic support, the hydraulic support comprising a traveling mechanism, a top beam, a hydraulic column, and a sensing system, the sensing system comprising a pose detection module and an environmental sensing module, characterized in that, Includes the following steps: The system acquires the tunnel environment information collected by the environmental perception module, extracts tunnel features based on the tunnel environment information, and initializes and calibrates the initial attitude data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support. The initial pose information is then used as the starting reference for the pose detection module to perform dead reckoning. Based on the initial pose information, the hydraulic column is controlled to release the support of the roof, and the traveling mechanism is driven to move the hydraulic support along the roadway. During the movement, the environmental perception module collects tunnel environmental perception data in real time and extracts tunnel environmental features based on the tunnel environmental perception data. The pose detection module performs dead reckoning starting from the initial reference to obtain the dead reckoning position; the cumulative error of the dead reckoning position is corrected using the tunnel environment features to obtain the corrected positioning coordinates. The hydraulic support is stopped according to the corrected positioning coordinates, and the hydraulic column is raised after stopping to complete the support of the top plate.

2. The control method for the hydraulic support according to claim 1, characterized in that, The tunnel environment information includes any one or more of the following: tunnel preset signage image, tunnel sidewall outline, tunnel cross-sectional shape, or fixed equipment features; the tunnel features include any one or more of the following: visual features, geometric contour features, or radar reflection features.

3. The control method for the hydraulic support according to claim 1, characterized in that, The initialization and calibration of the initial pose data detected by the pose detection module based on the tunnel features, to determine the initial pose information of the hydraulic support, includes: Determine the absolute coordinates of at least one reference target in the tunnel features in the tunnel coordinate system; Obtain the initial pose data detected by the pose detection module; Based on the absolute coordinates of the reference target and the initial attitude data detected by the pose detection module, the initial position coordinates and initial heading angle of the hydraulic support in the roadway coordinate system are calculated, and the initial position coordinates and initial heading angle are used as the initial pose information.

4. The control method for the hydraulic support according to claim 1, characterized in that, The method further includes: During the movement, safety assessments are made in real time based on the environmental perception data of the tunnel collected by the environmental perception module. Based on the safety assessment results, the hydraulic support is controlled to execute corresponding obstacle avoidance strategies to ensure the safety of the support operation.

5. The control method for the hydraulic support according to claim 4, characterized in that, The environmental perception module includes a ranging sensor, a camera, and an infrared detection device.

6. The control method for the hydraulic support according to claim 5, characterized in that, The ranging sensor is a millimeter-wave radar, which is deployed in at least four directions: front, rear, left, and right of the hydraulic support; the camera is deployed in at least four directions: front, rear, left, and right of the hydraulic support; and the infrared detection device is deployed in at least the front and rear directions of the hydraulic support.

7. The control method for the hydraulic support according to claim 6, characterized in that, The safety determination based on the roadway environmental perception data collected by the environmental perception module includes: Visual images are simultaneously acquired by the camera installed on the hydraulic support, infrared thermal images are simultaneously acquired by the infrared detection device, and the surrounding environment is continuously scanned by the millimeter-wave radar to obtain millimeter-wave radar ranging data. Obstacle features are determined based on the visual image, and the obstacle features are corrected based on the millimeter-wave radar ranging data to obtain the true physical features of the obstacle; Based on the infrared thermal imaging and the visual images, an intrusion determination of personnel into the dangerous area is made; A safety assessment is made based on the actual physical characteristics of the obstacle and the results of the personnel intrusion assessment of the dangerous area.

8. The control method for the hydraulic support according to claim 7, characterized in that, Determining obstacle features based on the visual image includes: Lens distortion correction is performed on the visual images captured by each camera based on the camera intrinsic parameters of each camera to obtain the corrected visual image. Feature points of the overlapping region of the corrected visual images corresponding to adjacent cameras are extracted using a feature matching algorithm. The homography matrix is ​​solved based on the feature points. The corrected visual images are then registered based on the homography matrix to obtain the image to be stitched together. A pixel-weighted fusion algorithm is used to stitch the images to be stitched together to eliminate stitching gaps and brightness abrupt changes, generating a 360° panoramic view image of the hydraulic support. Obstacle features in the 360° panoramic view image are extracted using an edge detection algorithm.

9. The control method for the hydraulic support according to claim 7, characterized in that, The step of correcting the obstacle features based on the millimeter-wave radar ranging data to obtain the true physical features of the obstacle includes: Based on the millimeter-wave radar ranging data and pixel spatial coordinate mapping algorithm, and according to the camera's internal and external parameter calibration parameters, the obstacle features are transformed from pixel coordinate system to underground physical coordinate system to obtain the true physical features of the obstacle.

10. The control method for the hydraulic support according to claim 8, characterized in that, The step of determining whether personnel have intruded into a dangerous area based on the infrared thermal image and the visual image includes: Based on a temperature threshold, pixels in the infrared thermal image are filtered to obtain suspected human targets in the infrared thermal image. The suspected human targets are screened using a connected component morphological feature discrimination mechanism to obtain the human target screening results; The human target recognition result is obtained by identifying human targets in the danger zone of the 360° panoramic view image using machine vision algorithms. Based on the results of human target screening and human target recognition, a determination of human intrusion into dangerous areas is made.

11. The control method for the hydraulic support according to claim 1 or 4, characterized in that, The step of performing dead reckoning using the pose detection module with the starting reference as the starting point to obtain the dead reckoning position includes: The initial position coordinates in the initial pose information are used as the integration starting point, and the initial heading angle in the initial pose information is used as the initial heading. The posture detection module collects the acceleration data of the hydraulic support along the roadway direction in real time. The acceleration data is integrated twice to obtain the displacement increment of the hydraulic support relative to the initial reference. The displacement increment is then added to the initial position coordinates to obtain the real-time calculated position coordinates.

12. The control method for the hydraulic support according to claim 5, characterized in that, The step of correcting the cumulative error of the dead reckoning position using the tunnel environment characteristics to obtain the corrected positioning coordinates includes: Based on the geometric contour features and / or spatial distribution features in the roadway environment, and combined with the distance data between the hydraulic support and the roadway sidewall collected in real time by the ranging sensor, the first absolute position of the hydraulic support in the transverse direction of the roadway is calculated. Based on the visual features in the tunnel environment, the preset support mileage scale or spacing mark in the tunnel is identified by the environmental image captured by the camera, and the second absolute position of the hydraulic support in the longitudinal direction of the tunnel is obtained. The lateral coordinates of the dead reckoning position are weighted and fused with the first absolute position to obtain the corrected lateral positioning coordinates; The longitudinal coordinates of the dead reckoning position are weighted and fused with the second absolute position to obtain the corrected longitudinal positioning coordinates; The corrected lateral positioning coordinates and the corrected longitudinal positioning coordinates are combined to form the corrected positioning coordinates.

13. A control system for a hydraulic support, the hydraulic support comprising a traveling mechanism, a top beam, and a hydraulic column, characterized in that, The control system includes: The sensing device includes a pose detection module and an environmental sensing module. The environmental sensing module is used to collect roadway environmental information and to collect roadway environmental sensing data during the movement of the hydraulic support. The pose detection module is used to detect the attitude data of the hydraulic support and perform dead reckoning. The controller, which is signal-connected to the sensing device, the walking mechanism, and the hydraulic column, is configured to perform the following steps: The system acquires the tunnel environment information collected by the environmental perception module, extracts tunnel features based on the tunnel environment information, and initializes and calibrates the initial attitude data detected by the pose detection module based on the tunnel features to determine the initial pose information of the hydraulic support. The initial pose information is then used as the starting reference for the pose detection module to perform dead reckoning. Based on the initial pose information, the hydraulic column is controlled to release the support of the roof, and the traveling mechanism is driven to move the hydraulic support along the roadway. During the movement, the environmental perception module collects tunnel environmental perception data in real time and extracts tunnel environmental features based on the tunnel environmental perception data. The pose detection module performs dead reckoning starting from the initial reference to obtain the dead reckoning position; the cumulative error of the dead reckoning position is corrected using the tunnel environment features to obtain the corrected positioning coordinates. The hydraulic support is stopped according to the corrected positioning coordinates, and the hydraulic column is raised after stopping to complete the support of the top plate.

Images