A method and system for motion travel inversion and collaborative straightening based on multi-source sensing on coal mining equipment
Patent Information
- Application Number
- CN202611223108.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-12
- Publication Date
- 2026-09-15
Smart Images
Figure CN122752011A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent control technology for fully mechanized coal mining, specifically to a method and system for inverting and coordinating the shift travel of coal mining machinery based on multi-source sensing. Background Technology
[0002] In underground coal mining operations, the longwall face mainly consists of three supporting equipment: a coal mining machine, a scraper conveyor, and hydraulic supports. The hydraulic supports are used to support the roof to provide a safe working space. The scraper conveyor is responsible for transporting coal and serves as the running track for the coal mining machine. The coal mining machine travels back and forth along the scraper conveyor track to perform cutting operations. As the working face continues to advance, the pushing jacks at the bottom of the hydraulic supports are used to sequentially complete the pushing and pulling actions, so that the hydraulic supports and scraper conveyor move forward as a whole. The straightness control of the working face equipment arrangement during this pushing process is a fundamental link to ensure the continuous safe operation of the longwall mining equipment.
[0003] Existing face straightening technology mainly relies on directly configuring stroke sensors inside or outside the pushing jacks of each hydraulic support. In practical applications, the control system reads the values of these stroke sensors that are scattered on each support to obtain local relative displacement data between the support and the scraper conveyor. After the coal mining machine completes a cut, the system summarizes the stroke sensor data of hundreds of hydraulic supports on the entire face to calculate the equipment position and sends action commands to the electro-hydraulic control system of each hydraulic support to drive the jacks to extend and retract, so as to maintain the straight shape of the scraper conveyor and the hydraulic support.
[0004] However, the environment of underground fully mechanized mining faces in coal mines is characterized by high concentrations of dust, water vapor, and mechanical collisions and vibrations. The hardware mode of distributing a large number of stroke sensors on hydraulic supports is prone to component damage, resulting in a high overall equipment failure rate. At the same time, this direct measurement method can only obtain local relative displacement, and the measurement error will continue to accumulate as the working face advances. This will prevent the system from obtaining reliable absolute spatial position and orientation information of the working face, and consequently, the control system will find it difficult to achieve precise straightening control of the working face due to the accumulated deviation of the input data. Summary of the Invention
[0005] To address the shortcomings of existing technologies, this invention provides a method and system for multi-source sensing on-board coal mining face propulsion stroke inversion and coordinated straightening. This solves the problem that existing fully mechanized mining face straightening technologies rely on a large number of stroke sensors distributed across various hydraulic supports. This hardware approach is prone to component damage under harsh working conditions, leading to a high overall equipment failure rate. Furthermore, this sensor-dependent direct measurement method can only acquire relative displacement data of local equipment, and the measurement error accumulates as the working face advances. This results in the system being unable to obtain reliable absolute spatial pose information of the entire working face, ultimately causing the control system to struggle to achieve accurate straightening of the working face equipment shape due to accumulated deviations in the input data.
[0006] The first aspect of this invention provides a method for inverting and coordinating the shift travel of coal mining equipment using multi-source sensing, comprising the following sequential steps: Two-dimensional image data and three-dimensional point cloud data of hydraulic supports are acquired by multi-source sensors rigidly installed on the coal mining machine. The multi-source sensors are integrated sensing devices including cameras, three-dimensional lidar and inertial measurement units. The continuous pose of the coal mining machine body relative to the absolute coordinate system of the working face is obtained by combining continuous trajectory propagation and correction calculation. Semantic segmentation is performed on the two-dimensional image data of the hydraulic support and it is mapped and separated into the region of interest of the three-dimensional point cloud. In each region of interest of the three-dimensional point cloud, cylindrical surface and plane fitting are performed to extract the end face reference points. The hinge point between the column and the base is calculated using the constant length of the fixed cylinder and the local coordinate system of the support base is constructed. Based on the meshing of the coal mining machine's traveling wheels and pins, the contact between the guide slipper and the guide surface, and the support slipper and the trough support relationship, the continuous pose of the coal mining machine body is converted into the local coordinate system of the scraper conveyor in the current middle trough. The hinge point between the column and the base is calculated using the end face reference point, the column axis and the constant length of the fixed cylinder. The side push hinge point of the support is calculated based on the key surface of the base and the known fixed connection relationship. The side connection point of the conveyor is calculated using the local coordinate system of the scraper conveyor and the fixed structural dimensions. Then, the relative connection vector between the support and the scraper conveyor and the current distance and signed distance deviation along the push direction are calculated. The actual current pushing distance is inverted by the inverse kinematics of the pushing mechanism, and the target pushing distance is calculated by combining the signed distance deviation. The actual support curve is formed by connecting the hinge points of the columns and bases of each frame in series, and the actual conveyor curve is formed by connecting the conveyor side of each frame. The actual support curve, the actual conveyor curve, the target line for pulling the frame, and the target line for pushing the frame are compared. Through a two-stage collaborative decision-making process of pulling the frame first and then pushing the frame, the amount of frame to be pulled for each frame is output first. After the status is updated, the amount of frame to be pushed for each frame is output, and feedback correction is performed in the next coal mining cycle.
[0007] Preferably, the step of acquiring two-dimensional image data and three-dimensional point cloud data of the hydraulic support through multi-source sensors rigidly mounted on the coal mining machine, and obtaining the continuous pose of the coal mining machine body relative to the absolute coordinate system of the working face by combining continuous trajectory propagation and correction calculations, specifically includes: Complete the external parameter calibration, time synchronization, and installation posture calibration between the camera, 3D lidar, inertial measurement unit, and coal mining machine body; The high-frequency output data of the inertial measurement unit is pre-integrated to continuously propagate attitude, velocity, and position. This data is then used in conjunction with visual odometry and laser odometry for trajectory correction to obtain the continuous pose of the coal mining machine.
[0008] Preferably, the step of semantically segmenting and mapping the two-dimensional image data of the hydraulic support to a three-dimensional point cloud region of interest, extracting end face reference points by fitting cylindrical and planar surfaces within each three-dimensional point cloud region of interest, and calculating the hinge point between the column and the base using a fixed cylinder with a constant length specifically includes: Component-level semantic segmentation is performed on the two-dimensional image of a single hydraulic support to extract semantic pixel regions containing the column fixing cylinder, piston or piston rod, top beam and base. The semantic pixel regions are mapped to the three-dimensional point cloud data by calling the extrinsic parameter matrix and the imaging projection relationship model of the camera, and the independent three-dimensional point cloud regions of interest are separated. Cylindrical surface fitting is performed on the region of interest of the 3D point cloud corresponding to the column to calculate and obtain the spatial vector of the central axis of the column and the value of the cylinder diameter. The point cloud data distributed along the column axis obtained by the solution is segmented into cylindrical surfaces for fitting. The location where the diameter value changes spatially is identified. The diameter change feature is compared with the model prior feature recorded in the mechanism parameter library. The diameter change location that meets the model prior feature is determined as the section at the junction of the two cylindrical surfaces of the fixed cylinder and the piston or piston rod. The center point of the section at the junction of the two cylindrical surfaces is established as the end face reference point. A fixed axial distance value is set from the end face reference point to the hinge point between the column and the base. The fixed axial distance value is obtained from the system's preset mechanism parameter library. Based on the physical characteristics that the axial length of the fixed cylinder and the end structure size are constant, the position coordinates of the hinge point between the column and the base are calculated.
[0009] Preferably, the step of converting the continuous pose of the coal mining machine body into the local coordinate system of the scraper conveyor in the current middle trough based on the meshing of the traveling wheels and pins, the contact between the guide slipper and the guide surface, and the support relationship between the support slipper and the trough side, specifically includes: The state of the coal mining machine’s traveling wheels engaging with the scraper conveyor’s pins or toothed rails, the coal mining machine’s guide slippers being restricted by the pins’ guide grooves or corresponding guide surfaces, and the coal mining machine’s support slippers being supported on the conveyor’s trough or shovel plate. Based on the known installation dimensions and fixed transformations of the guide pair between the coal mining machine and the scraper conveyor, the current meshing and contact state, and clearance compensation, the local coordinate system of the scraper conveyor corresponding to the current position of the coal mining machine is derived. Based on the mileage of the coal mining machine along the working face, the center distance of the hydraulic supports, the observation time, and the nearest neighbor projection, each local pose sample of the conveyor is matched with the corresponding hydraulic support to obtain the local pose sequence of the scraper conveyor under the trajectory constraint of the coal mining machine.
[0010] Preferably, the step of calculating the support-side pushing hinge point based on the key surface of the base and the known fixed connection relationship, calculating the conveyor-side connection point based on the local coordinate system of the scraper conveyor and the fixed structural dimensions, and then calculating the relative connection vector between the support and the scraper conveyor, as well as the current distance and the signed distance deviation along the pushing direction, specifically includes: Based on the coordinates of the hinge point between the column and the base, the spatial parameters of the key surface of the base, and the known fixed dimensions in the mechanism parameter library, the local coordinates of the side-pushing hinge point of the support and other fixed reference points are derived and calculated, and then uniformly converted to the absolute coordinate system of the working face through the continuous fusion trajectory data of the coal mining machine. The current closed-loop geometry of the mechanism is defined by the vector from the hinge point on the support side to the connection point on the conveyor side, and the current distance along the pushing direction is calculated by projecting the relative connection vector onto the pushing direction. The signed distance deviation is calculated by comparing the predetermined target relative distance with the current distance along the pushing direction, and the signed distance deviation is used to distinguish between underpush and overpush states.
[0011] Preferably, the step of inverting the current actual displacement distance through the inverse kinematics of the displacement mechanism and calculating the target displacement distance in combination with the signed distance deviation specifically includes: The hinge points on the support side, the connection points on the conveyor side, the hinge points at both ends of the pushing jack, the length of the pushing rod, and the dimensions of other connecting rods are combined to form a closed chain of mechanisms configured according to the model. Based on the current coordinates of the connection points on both sides and the mechanism parameters, the current actual pushing stroke is inverted through the inverse kinematics of the mechanism. Based on the signed distance deviation, the conveyor-side connection point is updated to the target geometric position along the specified pushing direction, and the target pushing stroke is calculated using the same mechanism inverse kinematics model; The required displacement compensation amount is calculated based on the target displacement stroke and the actual displacement stroke. A limiting operation is performed using the limiting range parameters determined by the cylinder stroke boundary, speed boundary, safety boundary and current process state, and the executable displacement amount is output.
[0012] Preferably, the comparison of the actual curve of the support, the actual curve of the conveyor, and the target lines for pulling and pushing the supports, through a two-stage collaborative decision-making process of pulling the supports first and then pushing them, first outputs the amount of supports to be pulled for each support. The first stage of the two-stage collaborative control specifically includes the following: The actual curve of the support frame and the target line of the support frame are calculated separately, and there is a sign deviation in the support frame stage. Preset the allowable deviation of the tie rod to construct the tie rod set to be adjusted; Calculate the local response coefficient between the actual unit direction vector of the support frame and the signed unit normal vector of the support frame target line at the corresponding position. Calculate the support frame distance to be pulled based on the signed deviation and directional projection relationship of each hydraulic support frame, and output the support frame amount to be pulled for each support frame.
[0013] Preferably, the first stage of the breakup decision-making process further includes: When the local response coefficient is detected to be less than the preset angle threshold, the tension distance is not directly output. Instead, the Jacobian conversion of the complete mechanism is used, the local direction of the target line is adjusted, or the corresponding frame control output is frozen. After the scaffolding action is completed, the scaffolding data is used to calculate and update the local coordinate system of the hinge point between the column and the base and the support base, and then the corresponding support enters the second stage of the pushing control program.
[0014] Preferably, the second-stage shift decision of the two-stage coordinated control specifically includes: Preset the allowable deviation for displacement and the allowable deviation for the relative distance between the support and the conveyor, and construct a set of displacement adjustments to be made; Using the actual conveyor curve, target line, current relative distance, and actual travel distance calculated after the status update as input sources, the required travel distance for each support frame is determined. The cost function for constructing the control quantity to be issued is optimized for smoothness, and the cost function is constrained by data weight coefficient, local geometric response coefficient and smoothness penalty coefficient. Under the premise of meeting the single-frame hydraulic cylinder stroke, execution speed, mechanism connection angle, top plate condition restrictions, and safety interlock anti-collision requirements, the output shows the sequence of actions to be performed for each frame and the amount of movement to be made for each frame.
[0015] A second aspect of the present invention provides a coal mining machine-mounted multi-source sensing shift stroke inversion and coordinated straightening system, which utilizes a control host to execute the above-described method, including: The system hardware platform includes a camera, a three-dimensional lidar, and an inertial measurement unit that are fixed to the same shell or rigid base and rigidly installed on the coal mining machine, constituting the integrated sensing device; The control logic unit, integrated within the control host, includes a time synchronization and calibration unit, a coal mining machine trajectory fusion unit, a hydraulic support semantic segmentation unit, a point cloud region of interest and geometric solution unit, a coal mining machine and conveyor constraint solution unit, a connection point geometric calculation unit, a push stroke inversion unit, a two-type curve and target line construction unit, a push-pull decision unit, and a control interface. The mechanism parameter library is used to record the corresponding column dimensions, constant axial length of the fixed cylinder, relative position of the hinge point, fixed dimensions of the guide shoe, and fixed connection parameters of the pushing linkage mechanism according to the model of the support, pushing mechanism, coal mining machine, and conveyor.
[0016] This invention provides a method and system for multi-source sensing-based displacement inversion and coordinated straightening in coal mining machinery. It offers the following advantages: 1. This invention rigidly mounts an integrated sensing device, including a camera, a 3D lidar, and an inertial measurement unit, onto a coal mining machine and combines continuous trajectory propagation and correction calculations to obtain continuous pose in the absolute coordinate system. This changes the hardware mode of distributing stroke sensors inside the hydraulic support and utilizes the movement characteristics of the coal mining machine to continuously collect full-domain 3D data. This achieves the measurement effect of reducing the probability of component failure and obtaining reliable global spatial pose information of the working face under harsh mining conditions.
[0017] 2. This invention extracts the end face reference points of the diameter-change section by segmenting cylindrical fitting of point cloud data, and deduces the coordinates of the hinge point between the column and the base by combining the physical constraint of the constant length of the fixed cylinder. Based on the mechanical contact state between the coal mining machine and the scraper conveyor, the pose of the coal mining machine is converted into the local pose of the conveyor. The coordinates of the connection point are input into the inverse kinematics model of the pushing mechanism to invert the actual pushing stroke. This achieves the effect of accurately obtaining the global absolute coordinates and the true pushing amount of the equipment without the influence of local field of view occlusion.
[0018] 3. This invention constructs corresponding actual curves by connecting the support and the conveyor side connection points in series and comparing them with the preset target line. A state update link is set between the support pulling and pushing processes to perform a two-stage collaborative decision-making process of pulling the support first and then pushing. In the pushing decision, a cost function containing local geometric response coefficients and smoothing penalty coefficients is used to optimize the issued pushing amount. This achieves a smooth straightening effect that suppresses abrupt changes in the control amount of adjacent supports and prevents mechanical interference and collisions of equipment under physical boundary constraints. Attached Figure Description
[0019] Figure 1 This is a diagram showing the relationship between the coal mining machinery integrated sensing device and the hydraulic support observed in this invention. Figure 2 This is a diagram showing the definition of the three key points, coordinate system, and geometric calculation of the connection points in this invention. Figure 3This is an inversion diagram of the local pose and traverse stroke of the scraper conveyor from the trajectory of the coal mining machine according to the present invention; Figure 4 The actual curves of the support frame, the actual curves of the conveyor, and the target curve diagram of the present invention are shown below. Figure 5 This is the two-stage pull-pull decision and feedback control diagram of the present invention. Detailed Implementation
[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0021] Example: See attached document Figure 1 , Figure 1 This is a diagram showing the relationship between an integrated sensing device for coal mining machinery and a hydraulic support, according to an embodiment of the present invention. To realize a method for multi-source sensing of coal mining machinery to reverse the movement of the machine and coordinate straightening, this embodiment provides a system that integrates spatial sensing, pose inversion, and coordinated straightening control logic. The overall implementation plan of the system provided in this embodiment covers the construction of the system hardware architecture, the configuration of the mechanism parameter library, and the execution of the overall method flow.
[0022] Step S1 involves building a system hardware platform and control logic unit consisting of an integrated sensing device for coal mining machinery, and configuring a mechanism parameter library. The integrated sensing device includes a camera, a 3D LiDAR, and an inertial measurement unit. The aforementioned camera, 3D LiDAR, and inertial measurement unit are fixed on the same housing or rigid base and rigidly installed on the coal mining machine. The system integrates a time synchronization and calibration unit, a coal mining machine trajectory fusion unit, a hydraulic support semantic segmentation unit, a point cloud region of interest and geometric solution unit, a coal mining machine and conveyor constraint solution unit, a connection point geometric calculation unit, a push stroke inversion unit, a two-type curve and target line construction unit, a push-pull decision unit, and a control interface within the control host. The system establishes a mechanism parameter library, recording the corresponding column dimensions, constant axial length of the fixed cylinder, relative position of the hinge point, fixed dimensions of the guide shoe, and fixed parameters such as the push linkage mechanism according to the model of the support, push mechanism, coal mining machine, and conveyor.
[0023] See attached document Figure 5 , Figure 5 According to an embodiment of the present invention, a two-stage pull-and-push decision and feedback control diagram is used to complete online measurement, geometric inversion and control decision-making according to the following consecutive steps when the coal mining machine is running along the working face: Execute step S2 to complete the calibration of multi-source sensors and continuous positioning of the coal mining machine. Complete the external parameter calibration, time synchronization and installation posture calibration between the camera, 3D LiDAR, inertial measurement unit and coal mining machine body. Use the high-frequency output data of the inertial measurement unit for pre-integration to continuously propagate attitude, velocity and position, and use visual odometry and laser odometry for calibration to obtain the continuous posture of the coal mining machine.
[0024] In step S3, semantic segmentation and geometric feature extraction of the hydraulic support are performed. Instance segmentation and component-level semantic segmentation are performed on the image for each support. The semantic regions of the column fixing cylinder, piston or piston, top beam and base are mapped into three-dimensional point cloud regions of interest. Cylindrical surface and plane fitting are performed in each three-dimensional point cloud region of interest. The cylinder and piston end faces are located by the abrupt change in the diameter of the coaxial cylindrical surface. The hinge point between the column and the base and the local coordinate system of the hydraulic support base are calculated using the constant length of the fixed cylinder.
[0025] Execute step S4 to invert the local pose of the scraper conveyor. Based on the meshing of the coal mining machine's traveling wheels and pins, the contact between the guide slipper and the guide surface, and the support relationship between the support slipper and the trough side, the coal mining machine's pose is converted into the local coordinate system of the scraper conveyor in the current middle trough.
[0026] Execute step S5 to perform geometric calculations and relative distance calculations at the connection points. Calculate the hinge point between the column and the base using the end face position, column axis, and constant length of the fixed cylinder. Calculate the side-pushing hinge point of the support based on the key surface of the base and the known fixed connection relationship. Calculate the side connection point of the conveyor using the local coordinate system of the scraper conveyor and the fixed structural dimensions. Then calculate the relative vector between the support and the scraper conveyor, as well as the current distance along the pushing direction.
[0027] Execute step S6 to invert the pushing stroke and construct two types of actual curves. Invert the current actual pushing stroke through the inverse kinematics of the pushing mechanism; form the actual curve of the support from the support reference point and the actual curve of the conveyor from the connection point on the conveyor side.
[0028] In step S7, a collaborative push-pull decision is generated and feedback correction is performed. The actual curves of the supports and the actual curves of the conveyors are compared with the target lines for pulling and pushing the supports. The system first outputs the amount of pulling the supports one by one, and then outputs the amount of pushing the supports one by one after the status is updated. After the control command is executed, the coal mining machine performs another inspection to correct the position of the supports, the shape of the scraper conveyor, and the mechanism parameters in a closed loop.
[0029] See attached document Figure 2 , Figure 2 This is a diagram showing the definition of the three key points, the coordinate system, and the geometric calculation of the connection points according to an embodiment of the present invention.
[0030] In step S301, the system performs image semantic segmentation and point cloud region of interest mapping on the hydraulic support. The system simultaneously acquires two-dimensional image data collected by the camera and three-dimensional point cloud data collected by the three-dimensional LiDAR. The two-dimensional image data is segmented frame by frame to complete the boundary division and number association of different hydraulic supports. Based on the instance segmentation, the system performs component-level semantic segmentation on the two-dimensional image of a single hydraulic support to extract semantic pixel regions containing the column fixing cylinder, piston or piston rod, top beam and base.
[0031] By calling the camera and 3D LiDAR extrinsic parameter matrices and the camera imaging projection relationship model obtained in the multi-source sensor joint calibration stage, semantic pixel regions are mapped to synchronously acquired 3D point cloud data, and the independent 3D point cloud regions of interest corresponding to the physical space of each component are separated.
[0032] In step S302, geometric feature fitting is performed within the region of interest of the 3D point cloud. Cylindrical surface fitting is performed within the region of interest of the 3D point cloud corresponding to the column. The spatial vector of the central axis of the column and the diameter of the column are calculated and obtained. Plane fitting is performed within the regions of interest of the 3D point cloud corresponding to the top beam and the 3D point cloud corresponding to the base, respectively. The spatial plane equation parameters and normal vectors of the key surfaces of the top beam and the base are calculated and obtained.
[0033] In step S303, extract the diameter abrupt change features of the coaxial cylindrical surface and locate the end face reference point. Within the region of interest of the three-dimensional point cloud of the same column, perform segmented cylindrical surface fitting on the point cloud data distributed along the solved column axis. Compare the cylindrical surface diameter values obtained from the segmented fitting to identify the location where the diameter value undergoes a spatial abrupt change. Compare the diameter abrupt change features with the fixed cylinder design outer diameter and piston design outer diameter values recorded in the mechanism parameter library. Determine the diameter abrupt change location that conforms to the model prior features in the parameter library as the cross section at the junction of the two cylindrical surfaces of the fixed cylinder and the piston or piston rod.
[0034] The center point of the interface between the two cylindrical sections is established as the end face reference point. Combined with end face reference points The point cloud in the vicinity is fitted with local planar or circular boundaries, and the end face reference points are corrected. The three-dimensional coordinates and normal vector, end face reference point It is only used for subsequent calculation of column spatial positioning and geometric hinge points, and is not used to output the specific extension value of the column.
[0035] Execute step S304, calculate the hinge point between the column and the base based on the physical constant length constraint, and set... The value is obtained by fitting the cylindrical surface of the fixed cylinder and is set according to the unit axial vector specified in the direction of the base pointing towards the top beam. end face reference point To the hinge point between the column and the base Fixed axial distance value, fixed axial distance Retrieved from the system's pre-set mechanism parameter library.
[0036] Based on the physical characteristics that the axial length of the fixed cylinder and the end structural dimensions remain constant, in the sensor coordinate system, the hinge point between the column and the base... Position coordinates The calculation formula is:
[0037] In the formula, Indicates the end face reference point Three-dimensional coordinates in the sensor coordinate system.
[0038] Execute step S305 to construct the local coordinate system of the hydraulic support base and perform global pose transformation, combining the column axis and end face reference points. And the key surfaces of the base, construct and establish the local coordinate system of the support base. When encountering local field-of-view occlusion, the pose prior data obtained from the key surface normal of the top beam, the key surface normal of the base, and adjacent data frames are combined to eliminate the rotation uncertainty around the column axis.
[0039] Using the column axis, the base key surface normal, and multi-frame geometric fitting results, the local coordinate system of the support base is obtained by calculation. Pose transformation matrix relative to the absolute coordinate system of the working surface The system connects the column and the base at the hinge point. It is designated as the unified primary reference point for the subsequent construction of the actual curve of the stent.
[0040] Based on the hinge point between the column and the base The coordinates of the base, the spatial parameters of the key surface of the base, and the known fixed dimensions in the mechanism parameter library are used to derive and calculate the local coordinates of the support side push hinge point and other fixed reference points. The visible support feature points in the sensor coordinate system are then uniformly converted to the working face absolute coordinate system through the continuous fusion trajectory data of the coal mining machine.
[0041] Hinge point between column and base Unified global coordinates in the absolute coordinate system of the working surface The calculation formula is:
[0042] In the formula, Indicates the observation time The continuous pose transformation matrix of the coal mining machine body relative to the absolute coordinate system of the working face at all times; This represents the fixed extrinsic parameter transformation matrix of the combined sensor coordinate system relative to the coal mining machine body coordinate system; With step S304 Indicating the same feature, both represent the hinge point between the column and the base. Position coordinates in the sensor coordinate system.
[0043] In step S306, the system performs measurement data confidence determination and state anomaly filtering. If any of the following occurs during the calculation process: insufficient coverage of the region of interest in the column point cloud; the fitted diameter value of the column surface does not meet the model prior configured in the mechanism parameter library; the coaxial column surface diameter changes abruptly and the position is unstable; the fitting residual of the key surface of the top beam or the key surface of the base exceeds the limit value; the constant length constraint residual of the fixed cylinder exceeds the limit value; the support number association has a logical conflict; or the multi-frame fusion calculation residual exceeds the limit value, the system will mark the hydraulic support with the corresponding number as a low confidence state. The hydraulic support marked as a low confidence state will not participate in the current control quantity calculation. The system will perform state parameter update after subsequent detection of effective observation data and recovery to normal.
[0044] See attached document Figure 3 , Figure 3 This is an inversion diagram of the coal mining machine trajectory to the local pose and traverse of the scraper conveyor according to an embodiment of the present invention.
[0045] Execute step S401 to establish a three-machine mechanical constraint and mechanism parameterized model. When the coal mining machine runs on the scraper conveyor, the coal mining machine's traveling wheels mesh with the scraper conveyor's pins or toothed rails. The coal mining machine's guide slippers are restricted by the pin guide grooves or corresponding guide surfaces. The coal mining machine's supporting slippers are supported on the conveyor trough side or shovel plate. The aforementioned mechanical relationships cause the relative pose between the coal mining machine's body coordinate system and the current central trough local coordinate system to be determined by structural dimensions, meshing position, slipper contact state, and finite clearance. On one side of the hydraulic support, push the thousand The cylinder end of the jack is hinged to the support base. The other end of the jack is connected to the push rod. The front end of the push rod is then hinged to the connecting lug or connecting device in the middle groove. Different models of hydraulic supports may adopt a positive mounting form, an inverted mounting form, or different linkage forms. The positive and negative relationship between the extension of the jack and the forward movement of the conveyor is not uniform. The system does not preset a fixed extension direction. The system calls the mechanism topology, length, hinge point, zero position, and direction symbol of a specific model of hydraulic support through the parameter library, and sets the hydraulic support base to always remain horizontal in the solution model.
[0046] In step S402, the continuous trajectory propagation and correction calculation of the coal mining machine is performed. The inertial measurement unit outputs angular velocity data at high frequency, and the trajectory fusion unit performs offset compensation, gravity processing, and pre-integration. The attitude, velocity, and position of the combined sensors and the coal mining machine body are continuously propagated. The visual odometry and laser odometry use the hydraulic support and the stable background structure within the field of view for trajectory correction. The system uses the stable and recurring hydraulic support features as landmarks to suppress cumulative drift along the working face direction. The pose of the combined sensors is converted into the pose of the coal mining machine through a rigid mounting transformation matrix. The specific calculation formula is as follows:
[0047] In the formula, This represents the continuous pose transformation matrix of the coal mining machine body relative to the absolute coordinate system of the working face; This represents the continuous pose transformation matrix of the combined sensor relative to the absolute coordinate system of the working surface; This represents the fixed external parameter transformation matrix of the combined sensor coordinate system relative to the coal mining machine body coordinate system.
[0048] Execute step S403 to invert the local pose of the scraper conveyor. Based on the known installation dimensions and fixed transformations of the guide pair between the coal mining machine and the scraper conveyor, the current meshing and contact state, and clearance compensation, the local coordinate system of the scraper conveyor corresponding to the current position of the coal mining machine is derived. The inversion formula for the local pose of the scraper conveyor is:
[0049] In the formula, Indicates the observation time The pose transformation matrix of the local coordinate system of the scraper conveyor relative to the absolute coordinate system of the working surface at any time; The pose transformation matrix function represents the local coordinate system of the scraper conveyor relative to the coordinate system of the coal mining machine. The relative motion parameters between the coal mining machine and the scraper conveyor include the meshing phase of the traveling wheels, the contact state of the slipper, and the allowable small attitude changes. This represents the guide constraint error term, which includes guide clearance, equipment wear, operating vibration, and local center groove angle.
[0050] Execute step S404 to perform the matching and association between the local pose samples of the conveyor and the hydraulic supports. The system matches each local pose sample of the conveyor with the first local pose sample of the hydraulic supports according to the mileage of the coal mining machine along the working face direction, the center distance of the hydraulic supports, the observation time, and the nearest neighbor projection. The hydraulic support is used to obtain the local pose sequence of the scraper conveyor under the constraint of the coal mining machine trajectory, instead of directly treating the coal mining machine trajectory as the conveyor curve.
[0051] See attached document Figure 2 , Figure 2 This is a diagram showing the definition of the three key points, the coordinate system, and the geometric calculation of the connection points according to an embodiment of the present invention.
[0052] Execute step S501 to solve the key spatial connection points. After completing the calculation of the coordinate system of the support base and the local coordinate system of the conveyor, the system determines three calculation points based on the end face positioning results, the constant length constraint of the fixed cylinder, and the fixed dimensions of the mechanism: a unified support reference point. , support side push hinge point and the connection point on the conveyor side Unified stent reference point The reference point is calculated along the column axis from the cylinder end face reference point, based on the constant axial length of the fixed cylinder, and is defined as the hinge point between the column and the base. When a specific model of equipment uses an alternative reference point fixed to the base... At that time, the system will replace the reference point according to a known fixed transformation. Unified conversion to a unified stent reference point To distinguish between end face measurement, coordinate transformation, and fixed offset, each key point is calculated in the following order according to the formulas:
[0053]
[0054]
[0055]
[0056] In the formula, This indicates the coordinates of the end face reference point in the sensor coordinate system. The unit axial vector is represented by the end face reference point and the unit axial vector, which are obtained by fitting the planar or cylindrical surface of the semantic point cloud region of interest. This represents the fixed axial distance from the end face reference point to the hinge point. This represents the offset vector from the hinge point to the side-shifted hinge point of the bracket. The aforementioned fixed axial distance and offset vector are determined by the dimensions of the bracket's fixed structure. and The rotation matrix and translation vector represent the sensor pose, which is obtained from the coal mining machine trajectory and installation relationship; The rotation matrix represents the local coordinate system of the support base relative to the absolute coordinate system of the working surface; and The rotation matrix and translation vector represent the local pose of the scraper conveyor, which is obtained by inversion of the guidance relationship. This represents the fixed geometric coordinates of the conveyor-side connection point within the local coordinate system of the scraper conveyor; This represents the clearance compensation vector, which is determined by the mechanism clearance and historical calibration.
[0057] Execute step S502, calculate the relative connection vector and distance deviation between the support and the scraper conveyor. The system defines the current closed-chain geometric relationship of the mechanism using the vector from the hinge point on the support side to the connection point on the conveyor side, and projects the relative connection vector onto the first... The specified pushing direction within the bottom surface of the support frame relative connection vector and the current distance along the direction of movement The calculation formula is:
[0058]
[0059] The system pre-determines the first step based on the equipment design status, the completed status of the support frame, the planned target line, or the calibration test. The target relative distance between the support frame and the scraper conveyor along the specified pushing direction In the working condition where the minimum safe distance after the scaffolding is completed is used as the control benchmark, let In operating conditions where it is necessary to maintain the spacing between other processes, Take the corresponding planned value, the signed distance deviation of the current state relative to the target state. Defined as:
[0060] Signed distance deviation The symbol indicates the direction in which the connection point on the conveyor side should move along the specified pushing direction, and the absolute value represents the difference between the target geometric distance and the current geometric distance. The system also retains a three-dimensional vector. Current distance along the direction of movement The components and elevation difference information along the working surface are used to distinguish between under-push and over-push states using signed distance deviation; when a specific model uses the jack axis or a specified vertical line as the measurement reference, the system uses the measurement reference direction instead of the average push direction.
[0061] In step S503, based on the inverse kinematics of the mechanism, the actual pushing stroke is calculated. The system assembles the support side hinge point, the conveyor side connection point, the hinge points at both ends of the pushing jack, the length of the pushing rod, and other connecting rod dimensions into a closed chain of mechanisms configured according to the model. This is done based on the current coordinates of the connection points on both sides and the mechanism parameters. Through inverse kinematics of the mechanism, the first The current actual travel distance of the frame :
[0062] In the formula, This represents the inverse kinematic function of the specific model's matching conveyor mechanism. This represents the set of mechanism parameters, indicating the actual pushing stroke when the pushing mechanism is locally linearized within its rated working range. The following approximate formula is used for calculation:
[0063] In the formula, Indicates the installation direction symbol, distinguishing between upright installation, inverted installation, and the corresponding relationship between extension and retraction; This indicates the local transmission ratio converted from relative distance to cylinder stroke; Indicates the reference geometric distance; This indicates the installation zero position and idle travel compensation values; The attitude compensation item is determined by the push rod swing angle, the joint small rotation angle, and the local attitude of the central groove. "Current distance minus target distance" is only used to characterize the target geometric distance deviation and should not be directly used as the cylinder extension amount that is universal for all models without referring to the specific mechanism drawings.
[0064] Execute step S504 to calculate the target travel and displacement compensation amount, based on the signed distance deviation. The system will connect the conveyor-side connection point along the first... The specified pushing direction is updated to the target geometric position. The target stroke was calculated using the same inverse kinematics model of the mechanism. :
[0065]
[0066] The difference between the current distance and the target distance first forms the target's geometric displacement, which is then converted into the target cylinder stroke through the specific model's mechanism geometric chain. and actual travel distance Calculate the first Required displacement compensation :
[0067] In the formula, This represents the amplitude limiting operation function. and The limit range parameter is determined by the cylinder stroke boundary, speed boundary, safety boundary and current process status. The system repeats the above calculation process for each support frame and outputs the actual pushing status, under-pushing amount, over-pushing amount, executable pushing amount and calculation confidence. The independent pushing stroke sensor is not a necessary input. If an independent pushing stroke sensor is configured on site, it is only used for model calibration, redundancy verification and fault diagnosis.
[0068] See attached document Figure 4 and attached Figure 5 , Figure 4 These are actual curves of the support frame, actual curves of the conveyor, and target curve diagrams according to an embodiment of the present invention. Figure 5 This is a two-stage pull-pull decision and feedback control diagram according to an embodiment of the present invention.
[0069] Execute step S601 to construct two types of actual curve and target line models and calculate the straightness deviation. The system constructs two types of actual curve models with completely different data sources and control purposes according to the hydraulic support number sequence. The actual curve of the support... The hinge points between each column and the base are uniformly calculated using geometric methods. The three-dimensional coordinates are connected in series to reflect the physical shape of the support arrangement after the support process is completed; the actual curve of the conveyor. The conveyor-side connection point is obtained by inversion calculation based on the relationship between the coal mining machine trajectory, guiding constraints, and the fixed structure of the central trough. Composition, used to reflect the actual laying pattern of the central trench and the difference in displacement between adjacent supports, for alternative reference points selected due to model configuration or local observation conditions. It is necessary to first convert the known fixed connection relationship of the bracket to the hinge point between the column and the base. Then participate in the actual curve construction of the support. Different types, different sides or unconverted reference benchmarks are strictly prohibited from being used in the same working plane.
[0070] The system generates the target line for the support frame based on a robust fit of the design centerline, the location of the reference support, and the roadway control points or effective support points. The target moving line is generated based on the design of the conveyor centerline, the operating process requirements of the coal mining machine, the target position of the completed support frame, and the relative distance between the target support frame and the conveyor. For the curved sections of the working surface, the system uses piecewise straight lines, low-order splines, or preset design curves as target lines. The mathematical expressions for the two types of curves are as follows:
[0071]
[0072] Based on the two types of actual curves and their corresponding target lines, the signed deviations for the stretching stage and the pushing stage are calculated respectively:
[0073]
[0074] In the formula, The signed spatial distance calculation function from the actual point to the target line is used to calculate the overall straightness of the hydraulic support based on the signed deviation. and the overall straightness of the scraper conveyor :
[0075]
[0076] In step S602, the allowable deviation threshold is set and a set of supports to be adjusted is constructed. The coordinated straightening adopts a two-stage control decision of first pulling the supports and then pushing them. Between the two stages of pulling and pushing, the spatial state and hidden point position of the hydraulic supports are re-estimated. The system calculates the geometric deviation of each hydraulic support and sets the allowable deviation for pulling the supports. Allowable deviation of shift and the allowable deviation of the relative distance between the support and the conveyor Construct a set of support frames to be adjusted. and shift the set to be adjusted :
[0077]
[0078] Only when the first The system will only allow the action adjustment procedure to begin when the hydraulic support belongs to the corresponding set to be adjusted and the confidence level calculated by the aforementioned measurement steps reaches the set threshold.
[0079] Execute step S603, perform the first stage of scaffold collaborative decision-making and status update, the first stage decision is based on the actual curve of the scaffold. and the target line of the pull frame Set as the system input source For the first The actual unit direction vector of the support frame within the reference plane of the base bottom surface. For the target line of the pull frame, at the first Given the signed distance unit normal vector at the corresponding position of the frame, calculate the local response coefficient of the frame pulling action to the deviation of the target line normal. :
[0080] Based on the signed deviation of each hydraulic support And the distance of the support frame should be calculated based on the relationship between the direction projection. : when hour, ; when hour, ; In the formula, Represents the amplitude limiting function. and This represents the maximum permissible pull distance, where the local response coefficient is... When the angle is less than the preset threshold, it indicates that the curve deviation of the pull frame direction is difficult to be effectively eliminated along the current pull frame direction. The system does not directly output the aforementioned pull frame distance control quantity, but uses the complete mechanism Jacobian conversion, adjusts the local direction of the target line, or freezes the corresponding frame control output.
[0081] After the support frame pulling action is completed, the hinge point between the column and the base is calculated and updated using the command data sent during the pulling action. The local coordinate system of the support base is then used to initiate the second stage of the push control program for the corresponding support.
[0082] Execute step S604 to perform the second-stage collaborative decision-making and control quantity smoothing optimization. The second-stage decision is based on the actual conveyor curve calculated after the state update. , advance the target line Current relative distance and actual travel distance As the system input source, jointly determine the required displacement distance for each support frame. The actual amount of push control issued Distribute according to the set to be adjusted: when hour, ; when hour, ; To avoid abrupt mechanical changes in the control quantities of adjacent supports, the system performs smooth optimization on the cost function of the control quantities to be issued, while satisfying the constraints of single-support cylinder stroke, execution speed, mechanism connection angle, top plate conditions, and safety interlock collision prevention.
[0083] In the formula, This indicates the amount of support or displacement to be optimized. This represents the local geometric response coefficient for the corresponding control stage; the tension stage uses... The shifting stage is determined by the Jacobian matrix of the shifting mechanism or the local transmission proportional parameters; Represents the data weighting coefficients determined by the confidence level of perception measurements and the equipment anomaly level; This represents the smoothing penalty coefficient used to limit the difference in control volume between adjacent stents.
[0084] After smoothing and optimization, the final output parameters are: the amount of support to be pulled, the amount to be pushed, the action execution sequence, the control confidence level, and the prohibition of execution. Each hydraulic support executes the instructions one by one or in groups according to the preset sequence of the coal mining process, the safety interlock specifications, and the execution priority. The control logic of simultaneous action of all supports is not adopted. Only when there is a systematic overall deviation in the working face, causing most or all supports to exceed the allowable deviation threshold, does the system form a large-scale adjustment task.
[0085] Execute step S605, summarize the system's calculated data and output the collaborative straightening results. After completing the above full-process calculation and feedback control, the system outputs the following parameter variable sequence: continuous pose, motion trajectory, and mileage information along the working face direction of the coal mining machine and integrated sensing device in the working face coordinate system; three-dimensional coordinate set of each support column and base hinge point, support side pushing hinge point, and conveyor side connection point; corresponding support local coordinate system and conveyor local coordinate system; actual curve data composed of each reference point; maximum absolute deviation, root mean square deviation, and overall straightness. The following parameters are considered: degree, center distance between adjacent supports, adjacent front and rear misalignment, position difference between adjacent middle troughs, local bending angle of the conveyor, number of continuous abnormal deviation sections and supports exceeding the threshold; current relative distance, target relative distance, distance deviation, actual pushing stroke and target pushing stroke of each support pushing mechanism; target line for pulling supports, target line for pushing supports, set of supports to be adjusted for pulling supports and set of supports to be adjusted for pushing supports, as well as the amount of supports to be pulled, the amount to be pushed and the control execution sequence for each support; the calculated confidence level, abnormality level code, zero control quantity mark and prohibition mark corresponding to each measurement result and control quantity.
[0086] For the mathematical processing algorithm of the target line for low-order spline fitting and the driving action logic of the electro-hydraulic control valve group of the fully mechanized mining equipment after receiving the command, those skilled in the art can use existing numerical calculation methods and communication mechanisms of the mine electro-hydraulic control system. Low-order spline fitting and electro-hydraulic control logic are well-known technologies in this field and will not be elaborated here.
[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for coal mining machine load multi-source perception push travel inversion and collaborative straightening, characterized in that, Includes the following sequential steps: Two-dimensional image data and three-dimensional point cloud data of hydraulic supports are acquired by multi-source sensors rigidly installed on the coal mining machine. The multi-source sensors are integrated sensing devices including cameras, three-dimensional lidar and inertial measurement units. The continuous pose of the coal mining machine body relative to the absolute coordinate system of the working face is obtained by combining continuous trajectory propagation and correction calculation. Semantic segmentation is performed on the two-dimensional image data of the hydraulic support and it is mapped and separated into the region of interest of the three-dimensional point cloud. In each region of interest of the three-dimensional point cloud, cylindrical surface and plane fitting are performed to extract the end face reference points. The hinge point between the column and the base is calculated using the constant length of the fixed cylinder and the local coordinate system of the support base is constructed. Based on the meshing of the coal mining machine's traveling wheels and pins, the contact between the guide slipper and the guide surface, and the support slipper and the trough support relationship, the continuous pose of the coal mining machine body is converted into the local coordinate system of the scraper conveyor in the current middle trough. The hinge point between the column and the base is calculated using the end face reference point, the column axis and the constant length of the fixed cylinder. The side push hinge point of the support is calculated based on the key surface of the base and the known fixed connection relationship. The side connection point of the conveyor is calculated using the local coordinate system of the scraper conveyor and the fixed structural dimensions. Then, the relative connection vector between the support and the scraper conveyor and the current distance and signed distance deviation along the push direction are calculated. The actual current pushing distance is inverted by the inverse kinematics of the pushing mechanism, and the target pushing distance is calculated by combining the signed distance deviation. The actual support curve is formed by connecting the hinge points of the columns and bases of each frame in series, and the actual conveyor curve is formed by connecting the conveyor side of each frame. The actual support curve, the actual conveyor curve, the target line for pulling the frame, and the target line for pushing the frame are compared. Through a two-stage collaborative decision-making process of pulling the frame first and then pushing the frame, the amount of frame to be pulled for each frame is output first. After the status is updated, the amount of frame to be pushed for each frame is output. Feedback correction is performed in the next coal mining cycle.
2. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method according to claim 1, characterized in that, The process involves acquiring two-dimensional image data and three-dimensional point cloud data of the hydraulic support through multi-source sensors rigidly mounted on the coal mining machine, and combining this with continuous trajectory propagation and correction calculations to obtain the continuous pose of the coal mining machine body relative to the absolute coordinate system of the working face. Specifically, this includes: Complete the external parameter calibration, time synchronization, and installation posture calibration between the camera, 3D lidar, inertial measurement unit, and coal mining machine body; The high-frequency output data of the inertial measurement unit is pre-integrated to continuously propagate attitude, velocity, and position. This data is then used in conjunction with visual odometry and laser odometry for trajectory correction to obtain the continuous pose of the coal mining machine.
3. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method of claim 1, wherein, The process involves semantic segmentation and mapping of the two-dimensional image data of the hydraulic support to a three-dimensional point cloud region of interest. Within each region of interest, cylindrical and planar fitting is performed to extract end face reference points. The hinge point between the column and the base is calculated using a fixed cylinder with a constant length. Specifically, this includes: Component-level semantic segmentation is performed on the two-dimensional image of a single hydraulic support to extract semantic pixel regions containing the column fixing cylinder, piston or piston rod, top beam and base. The semantic pixel regions are mapped to the three-dimensional point cloud data by calling the extrinsic parameter matrix and the imaging projection relationship model of the camera, and the independent three-dimensional point cloud regions of interest are separated. Perform cylindrical surface fitting on the region of interest of the 3D point cloud corresponding to the column, and calculate and obtain the spatial vector of the central axis of the column and the value of the cylinder diameter. The point cloud data distributed along the column axis obtained by the solution is segmented into cylindrical surfaces for fitting. The location where the diameter value changes spatially is identified. The diameter change feature is compared with the model prior feature recorded in the mechanism parameter library. The diameter change location that meets the model prior feature is determined as the section at the junction of the two cylindrical surfaces of the fixed cylinder and the piston or piston rod. The center point of the section at the junction of the two cylindrical surfaces is established as the end face reference point. A fixed axial distance value is set from the end face reference point to the hinge point between the column and the base. The fixed axial distance value is obtained from the system's preset mechanism parameter library. Based on the physical characteristics that the axial length of the fixed cylinder and the end structure size are constant, the position coordinates of the hinge point between the column and the base are calculated.
4. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method of claim 1, wherein, Based on the meshing of the coal mining machine's traveling wheels and pins, the contact between the guide slipper and the guide surface, and the support slipper and the trough support relationship, the continuous pose of the coal mining machine body is converted into the local coordinate system of the scraper conveyor in the current central trough, specifically including: The state of the coal mining machine’s traveling wheels engaging with the scraper conveyor’s pins or toothed rails, the coal mining machine’s guide slippers being restricted by the pins’ guide grooves or corresponding guide surfaces, and the coal mining machine’s support slippers being supported on the conveyor’s trough or shovel plate. Based on the known installation dimensions and fixed transformations of the guide pair between the coal mining machine and the scraper conveyor, the current meshing and contact state, and clearance compensation, the local coordinate system of the scraper conveyor corresponding to the current position of the coal mining machine is derived. Based on the mileage of the coal mining machine along the working face, the center distance of the hydraulic supports, the observation time, and the nearest neighbor projection, each local pose sample of the conveyor is matched with the corresponding hydraulic support to obtain the local pose sequence of the scraper conveyor under the trajectory constraint of the coal mining machine.
5. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method of claim 1, wherein, The calculation of the support side-pushing hinge point based on the key surface of the base and the known fixed connection relationship, the calculation of the conveyor side connection point based on the local coordinate system of the scraper conveyor and the fixed structural dimensions, and then the calculation of the relative connection vector between the support and the scraper conveyor, as well as the current distance and the deviation of the signed distance along the pushing direction, specifically includes: Based on the coordinates of the hinge point between the column and the base, the spatial parameters of the key surface of the base, and the known fixed dimensions in the mechanism parameter library, the local coordinates of the side-pushing hinge point of the support and other fixed reference points are derived and calculated, and then uniformly converted to the absolute coordinate system of the working face through the continuous fusion trajectory data of the coal mining machine. The current closed-loop geometry of the mechanism is defined by the vector from the hinge point on the support side to the connection point on the conveyor side, and the current distance along the pushing direction is calculated by projecting the relative connection vector onto the pushing direction. The signed distance deviation is calculated by comparing the predetermined target relative distance with the current distance along the pushing direction, and the signed distance deviation is used to distinguish between under-pushing and over-pushing states.
6. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method according to claim 5, characterized in that, The process of inverting the current actual displacement distance through the inverse kinematics of the displacement mechanism and calculating the target displacement distance in combination with the signed distance deviation specifically includes: The hinge points on the support side, the connection points on the conveyor side, the hinge points at both ends of the pushing jack, the length of the pushing rod, and the dimensions of other connecting rods are combined to form a closed chain of mechanisms configured according to the model. Based on the current coordinates of the connection points on both sides and the mechanism parameters, the current actual pushing stroke is inverted through the inverse kinematics of the mechanism. Based on the signed distance deviation, the conveyor-side connection point is updated to the target geometric position along the specified pushing direction, and the target pushing stroke is calculated using the same mechanism inverse kinematics model; The required displacement compensation amount is calculated based on the target displacement stroke and the actual displacement stroke. A limiting operation is performed using the limiting range parameters determined by the cylinder stroke boundary, speed boundary, safety boundary and current process state, and the executable displacement amount is output.
7. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method of claim 1, wherein, The comparison of the actual curve of the support frame, the actual curve of the conveyor, and the target lines for pulling and pushing the support frame is used. Through a two-stage collaborative decision-making process of pulling the support frame first and then pushing it, the amount of support frame to be pulled for each frame is output first. The first stage of the two-stage collaborative control decision-making for pulling the support frame specifically includes: The actual curve of the support frame and the target line of the support frame are calculated separately, and there is a sign deviation in the support frame stage. Preset the allowable deviation of the tie rod to construct the tie rod set to be adjusted; Calculate the local response coefficient between the actual unit direction vector of the support frame and the signed unit normal vector of the support frame target line at the corresponding position. Calculate the support frame distance to be pulled based on the signed deviation and directional projection relationship of each hydraulic support frame, and output the support frame amount to be pulled for each support frame.
8. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method according to claim 7, characterized in that, The first phase of the decision-making process to break up the fight also includes: When the local response coefficient is detected to be less than the preset angle threshold, the tension distance is not directly output. Instead, the Jacobian conversion of the complete mechanism is used, the local direction of the target line is adjusted, or the corresponding frame control output is frozen. After the scaffolding action is completed, the scaffolding data is used to calculate and update the local coordinate system of the hinge point between the column and the base and the support base, and then the corresponding support enters the second stage of the pushing control program.
9. The coal cutter carried multi-source perception push travel inversion and collaborative straightening method according to claim 8, characterized in that, The second-stage shift decision for two-stage coordinated control specifically includes: Preset the allowable deviation for displacement and the allowable deviation for the relative distance between the support and the conveyor, and construct a set of displacement adjustments to be made; Using the actual conveyor curve, target line, current relative distance, and actual travel distance calculated after the status update as input sources, the required travel distance for each support frame is determined. The cost function for constructing the control quantity to be issued is optimized for smoothness, and the cost function is constrained by data weight coefficient, local geometric response coefficient and smoothness penalty coefficient. Under the premise of meeting the single-frame hydraulic cylinder stroke, execution speed, mechanism connection angle, top plate condition restrictions, and safety interlock anti-collision requirements, the output shows the sequence of actions to be performed for each frame and the amount of movement to be made for each frame.
10. A coal mining machine-mounted multi-source sensing system for shift stroke inversion and coordinated straightening, characterized in that, The method for inverting and coordinating the shift travel of a coal mining machine based on multi-source sensing as described in any one of claims 1 to 9, using a control host, includes: The system hardware platform includes a camera, a three-dimensional lidar, and an inertial measurement unit that are fixed to the same shell or rigid base and rigidly installed on the coal mining machine, constituting the integrated sensing device; The control logic unit, integrated within the control host, includes a time synchronization and calibration unit, a coal mining machine trajectory fusion unit, a hydraulic support semantic segmentation unit, a point cloud region of interest and geometric solution unit, a coal mining machine and conveyor constraint solution unit, a connection point geometric calculation unit, a push stroke inversion unit, a two-type curve and target line construction unit, a push-pull decision unit, and a control interface. The mechanism parameter library is used to record the corresponding column dimensions, constant axial length of the fixed cylinder, relative position of the hinge point, fixed dimensions of the guide shoe, and fixed connection parameters of the pushing linkage mechanism according to the model of the support, pushing mechanism, coal mining machine, and conveyor.