3D Scan-Based Arm Path Planning System for Tool Changer Robots in Limited Space

Abstract

This invention relates to the technical field of robotic arm adjustment, and discloses a 3D scanning-based arm span path planning system for a tool-changing robot within a limited space. The system includes a sensing module, a scanning module, and a path planning module. The sensing module identifies obstacle types and divides the robot's workspace into different sub-regions based on obstacle types. The scanning module acquires spatial scanning data for each sub-region using 3D scanning. The sensing module also selects reference feature points. The scanning module calibrates the spatial scanning data based on the reference feature points. The path planning module plans the robot's arm span path based on the spatial scanning data. This application improves the safety and efficiency of tool-changing operations through precise path planning.

Description

Technical Field

[0001] This invention relates to the technical field of robotic arm adjustment, and more specifically to a 3D scanning-based arm span path planning system for a tool-changing robot within a limited space. Background Technology

[0002] Cutterhead changing robots used in tunnel boring machines (TBMs) extend and move their end effectors within a confined space during cutterhead replacement, controlled by an arm extension control mechanism. Due to differences in TBM models and operating environments, the size and obstacles within the confined space for arm extension control vary. Current cutterhead changing robot technology still has many shortcomings.

[0003] The internal structure of a tunnel boring machine (TBM) is complex, with various mechanical components interwoven. Furthermore, during tunneling, it is subject to disturbances such as rockfalls and equipment vibrations, resulting in a dynamically changing working environment. Existing detection methods, such as simple sensor combinations or low-precision scanning equipment, are insufficient to comprehensively and accurately acquire information such as the location, shape, and type of obstacles. This makes it difficult for the cutter-changing robot to fully consider environmental factors when planning its path, increasing the risk of collisions.

[0004] Traditional path planning does not adequately incorporate the specific actions and constraints of tool changing operations for refined planning. In actual tool changing processes, different actions, such as tool grasping, transporting, and installing, place varying demands on the robot's arm's range of motion, posture, and force. Existing algorithms fail to optimize for these differences, resulting in less fluid robot movements during tool changing tasks, impacting the accuracy and efficiency of the tool changing process.

[0005] For example, Chinese patent application CN116551688A discloses a method for operating a robot for detecting or replacing cutterhead tools in a tunnel boring machine (TBM) and a TBM. The method includes: constructing a database of collision-free paths for the robot; obtaining the current posture and target posture of the robot arm; querying the database for collision-free paths from the current posture to the target posture and optimizing the paths; and controlling the robot arm to move from the current posture to the target posture along the optimized paths. This method effectively improves the robot's operating efficiency by calling paths from the database, but it still suffers from the problem mentioned in the background of this application: it does not fully incorporate the specific actions and constraints of the cutterhead replacement operation for refined planning.

[0006] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0007] The technical problem this invention aims to solve is to overcome the shortcomings of existing technologies and provide a 3D scanning-based arm path planning system for tool changing robots within a limited space. Through precise path planning, this system improves the safety and efficiency of tool changing operations. To solve the above technical problem, this invention provides the following technical solution:

[0008] A 3D scanning-based arm path planning system for a tool-changing robot within a limited space includes a perception module, a scanning module, and a path planning module; wherein:

[0009] The perception module is used to identify obstacle types; based on the type of obstacle, the perception module divides the workspace of the tool-changing robot into different sub-regions;

[0010] The scanning module acquires spatial scanning data for each sub-region based on 3D scanning;

[0011] The sensing module is also used to select reference feature points; the scanning module calibrates the spatial scanning data based on the reference feature points;

[0012] The path planning module performs arm span path planning for the tool-changing robot based on the spatial scanning data.

[0013] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the perception module includes an identification unit;

[0014] The identification unit is used to identify the types of obstacles in the tool-changing robot's workspace, specifically including:

[0015] The system acquires images of obstacles in the workspace; the recognition unit is equipped with a trained target detection model, and the obstacle images are input into the trained target detection model to identify the obstacle types.

[0016] The obstacle types include intrusion obstacles and structural obstacles; the intrusion obstacles include obstacles that accidentally intrude into the workspace of the tool-changing robot; the structural obstacles include the internal structure of the robot's work cabin.

[0017] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the perception module further includes an evaluation unit; the evaluation unit is used to divide the workspace of the tool-changing robot into different sub-regions, specifically including:

[0018] The coordinate range of each obstacle is obtained; the recognition unit is also equipped with a binocular vision device; the recognition unit calculates the coordinate range of each obstacle based on the binocular vision device and transmits it to the evaluation unit;

[0019] Based on the coordinate range and type of each obstacle, the workspace is divided into different types of sub-regions, and the coordinate range of each sub-region is recorded. The types of sub-regions include safe passage area, buffer zone, and core operation area. The core operation area is divided with the cutter box of the tunnel boring machine as the center. Outside the core operation area, the buffer zone is divided with any structural obstacle or intrusion obstacle as the center. In the workspace, the area outside the buffer zone and the core operation area is the safe passage area.

[0020] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space according to the present invention, the scanning module includes a scanning component unit; the scanning component unit is used to perform 3D scanning on each sub-region to obtain spatial scanning data of each sub-region, specifically including:

[0021] Set the scan resolution for each type of sub-region; read the type and coordinate range of each sub-region;

[0022] The workspace of the tool-changing robot is 3D scanned, and the coordinates of the scanned points are detected in real time.

[0023] The type of the sub-region where the scan point is located is identified based on the coordinates of the scan point, and the scan resolution is dynamically adjusted based on the type of the sub-region.

[0024] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the perception module further includes a feature selection unit; the scanning module further includes a calibration unit.

[0025] The feature selection unit is used to select reference feature points in the workspace of the tool-changing robot, specifically including: selecting no less than n reference feature points before the 3D scan begins; n is a positive integer; measuring the three-dimensional coordinates of each reference feature point and marking the actual coordinates of each reference feature point;

[0026] The calibration unit performs calibration and correction of spatial scan data based on reference feature points, specifically including:

[0027] Read the point cloud data composed of the spatial scan data; extract feature points from the point cloud data;

[0028] The extracted feature points are matched with the reference feature points to obtain the measured coordinates of each reference feature point in the point cloud data;

[0029] Based on the actual and measured coordinates of each reference feature point, calculate the rotation matrix and translation vector of the point cloud data;

[0030] The rotation matrix and translation vector are used to compensate and correct each spatial scan data in the point cloud data.

[0031] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the path planning module includes an action decomposition unit; the action decomposition unit is used to decompose the task of the tool-changing robot into multiple sub-actions, and combine spatial scanning data to determine the target state and constraints of each sub-action.

[0032] The target state includes the starting coordinates and ending coordinates of the end effector; the constraints include spatial constraints, attitude constraints, and force constraints.

[0033] As a preferred embodiment of the 3D scanning-based tool-changing robot arm span path planning system in a limited space as described in this invention, the path planning module further includes a path exploration unit; the path exploration unit explores alternative paths for each sub-action based on the target state and constraints of each sub-action, and generates a set of alternative paths for each sub-action; the set of alternative paths for any sub-action contains at least one alternative path for the corresponding sub-action.

[0034] The path exploration unit explores alternative paths for each sub-action based on the target state and constraints of each sub-action, specifically including:

[0035] Read the target state and obtain the start and end coordinates of the end effector of each sub-action;

[0036] Based on the path planning algorithm, considering the constraints, each sub-action plans at least M alternative paths; M is a positive integer; any alternative path for any sub-action is the movement of the end effector from the starting coordinate to the ending coordinate of the corresponding sub-action.

[0037] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the path planning module further includes a calculation unit; the calculation unit generates an arm span path based on the candidate path set for each sub-action; specifically including:

[0038] The calculation unit selects one alternative path from the set of alternative paths for each sub-action, and combines all the selected alternative paths for the sub-actions into a pending arm span path.

[0039] Perform a stability test on any two adjacent sub-movements in the undetermined arm span path; if all two adjacent sub-movements in the undetermined arm span path pass the stability test, then mark the undetermined arm span path as an arm span path.

[0040] The computing unit generates no fewer than m arm span paths and calculates the risk level of each arm span path; the arm span path with the lowest risk level is output as the arm span path planning result of the tool changing robot.

[0041] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the stability detection is specifically as follows:

[0042] Calculate the rate of change of acceleration of each joint of the robot in two adjacent sub-actions; the calculation unit is configured with an acceleration rate of change threshold; if the rate of change of acceleration of each joint is less than the acceleration rate of change threshold, then the two adjacent sub-actions pass the stability test.

[0043] As a preferred embodiment of the 3D scanning-based arm span path planning system for a tool-changing robot within a limited space as described in this invention, the computing unit calculates the risk level of each arm span path, specifically including:

[0044] Calculate the lengths of each alternative path included in any arm span path and sum them to obtain the total arm span path length;

[0045] Based on the alternative paths for each sub-action in the arm span path, calculate the motion of each joint of the tool-changing robot when executing each sub-action; based on the motion of each joint, calculate the motion trajectory of each joint.

[0046] Read the coordinate range of each buffer and the core working area in the workspace; calculate the total number of times all joints pass through the buffer and the core working area based on the motion trajectory of each joint, and record the total number of times as the risk count;

[0047] The risk level of the arm span path is obtained by normalizing the total length of the arm span path and the number of risks, and then weighting and summing them.

[0048] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0049] This application effectively avoids the risk of collisions with various obstacles during the operation of the tool changing robot by sensing and analyzing obstacles in the workspace of the tool changing robot, thus ensuring the safety of the tool changing operation and reducing the probability of equipment damage.

[0050] This application fully considers the complex requirements of tool changing operations, taking into account the spatial, posture, and force requirements of arm movement, and achieves highly refined path planning, making tool changing operations smoother and more efficient.

[0051] By rigorously calibrating and correcting the scanned data from the workspace, the impact of factors such as internal vibration of the tunnel boring machine and interference from complex environments on data accuracy was effectively overcome. This provides accurate and reliable data support for path planning. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments 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. Wherein:

[0053] Figure 1 A schematic diagram of the arm span path planning system for a tool-changing robot based on 3D scanning within a limited space provided by the present invention;

[0054] Figure 2 A flowchart illustrating the calibration and correction of spatial scanning data using the calibration unit provided in this invention;

[0055] Figure 3 A flowchart illustrating how the scanning component unit provided by the present invention acquires spatial scanning data for each sub-region;

[0056] Figure 4 A flowchart illustrating the calculation of the risk level of each arm span path by the computing unit provided in this invention. Detailed Implementation

[0057] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0058] This embodiment describes a 3D scanning-based arm path planning system for a tool-changing robot within a limited space, referring to... Figure 1 The system includes a perception module, a scanning module, and a path planning module; among which:

[0059] The perception module is used to identify obstacle types; based on the type of obstacle, the perception module divides the workspace of the tool-changing robot into different sub-regions;

[0060] The perception module includes an identification unit and an evaluation unit;

[0061] The identification unit is used to identify the types of obstacles in the tool-changing robot's workspace, specifically including:

[0062] Images of obstacles in the workspace are acquired. The recognition unit is equipped with a trained target detection model. The obstacle images are input into the trained target detection model to identify the obstacle type. For example, target detection algorithms such as Faster R-CNN and YOLO series can be used to accurately identify different obstacles. The model is trained specifically for various structures inside different tunnel boring machines, such as curved cutterhead supports and irregularly distributed grouting pipes. This allows the target detection model to quickly and accurately identify these structures, avoiding misidentification as passable spaces. This more accurately determines the actual movable space range of the boom mechanism, reduces scanning errors, and improves the safety of the cutterhead changing robot operation.

[0063] The types of obstacles include intrusion obstacles and structural obstacles; intrusion obstacles include obstacles that accidentally intrude into the workspace of the tool-changing robot; structural obstacles include the internal structure of the robot's work cabin, such as hydraulic pipes, hatches, etc.

[0064] The evaluation unit is used to divide the workspace of the tool-changing robot into different sub-regions, specifically including:

[0065] The recognition unit is equipped with a binocular vision device. The recognition unit calculates the coordinate range of each obstacle based on the binocular vision device and transmits it to the evaluation unit. The binocular vision device uses different images of the same obstacle captured by two cameras, and finds corresponding feature points in different images through a stereo matching algorithm. Then, it calculates the positional difference between these corresponding points and infers the three-dimensional coordinates of the feature points based on the positional difference. As a preferred embodiment of this application, edge detection is first performed on the obstacle image to extract the edge information of the obstacle. Then, a contour extraction algorithm is used to connect the obstacle edges to form the outline of the obstacle, thereby determining the approximate shape boundary of the obstacle. The coordinates of feature points on the obstacle outline are obtained based on the binocular vision device, thereby determining the coordinate range of each obstacle.

[0066] Based on the coordinate range and type of each obstacle, the workspace is divided into different types of sub-regions, and the coordinate range of each sub-region is recorded. The types of sub-regions include safe passage areas, buffer zones, and core operation areas. The core operation area is defined with the cutterhead of the tunnel boring machine (TBM) as the center. Outside the core operation area, buffer zones are defined with any structural obstacle or intrusion obstacle as the center. The area outside the buffer zones and core operation areas within the workspace is the safe passage area. For example, in this embodiment, a spherical or cuboid area extending uniformly outwards by 3 meters in all directions from the cutterhead as the center is preferably defined as the core operation area. The core operation area is the area directly operated for cutterhead replacement, where the installation and removal of cutters and bolts are all carried out. Further, in this embodiment, a buffer zone is also preferably defined with the rock strata as the center and extending uniformly outwards by 3 meters in all directions, or with the TBM's hatch as the center and extending uniformly outwards by 0.5 meters in all directions as the center. In subsequent spatial scanning and arm span path planning, special attention will be paid to the core working area and buffer zone to avoid collisions between the tool-changing robot and the tool it holds and obstacles.

[0067] The scanning module acquires spatial scanning data for each sub-region based on 3D scanning;

[0068] The scanning module includes a scanning component unit; refer to Figure 3 The scanning component unit is used to perform 3D scanning on each of the sub-regions to obtain spatial scanning data for each sub-region, specifically including:

[0069] Set the scan resolution for each type of sub-region; read the type and coordinate range of each sub-region;

[0070] The workspace of the tool-changing robot is 3D scanned, and the coordinates of the scanned points are detected in real time.

[0071] The type of the sub-region where the scan point is located is identified based on the coordinates of the scan point, and the scan resolution is dynamically adjusted based on the type of the sub-region.

[0072] In this application embodiment, the preferred scanning resolution is millimeter-level for the core work area, centimeter-level for the buffer zone, and decimeter-level for the safe passage area. For example, for the core work area, the scanning resolution is 5mm×5mm×5mm, meaning the voxel side length is 5mm; for any buffer zone, the scanning resolution is 3cm×3cm×3cm, meaning the voxel side length is 3cm; and for any safe passage area, the scanning resolution is 10cm×10cm×10cm, meaning the voxel side length is 10cm. When scanning the safe passage area, the scanning resolution of the scanning component is 10cm×10cm×10cm. If scanning the buffer zone is detected, the scanning resolution of the scanning component is switched to 3cm×3cm×3cm. This application employs a multi-resolution scanning strategy, using high-resolution scanning in the critical core work area to obtain detailed structural information, and using low-resolution scanning in the non-critical safe passage area to improve scanning speed; while ensuring scanning accuracy, it significantly shortens the overall scanning time.

[0073] The spatial scanning data includes the coordinates of all obstacles in the tool-changing robot's workspace, as well as the coordinates of each scanning point on the inner wall of the enclosed or semi-enclosed workspace. This coordinate data is stored in the form of point cloud data. Based on this point cloud data, the traversable spatial range of the tool-changing robot can be identified.

[0074] The sensing module is also used to select reference feature points; the scanning module calibrates the spatial scanning data based on the reference feature points;

[0075] The sensing module further includes a feature selection unit; the scanning module further includes a calibration unit;

[0076] The feature selection unit is used to select reference feature points in the workspace of the tool-changing robot, specifically including: selecting no less than n reference feature points before the 3D scan begins; n is a positive integer; measuring the three-dimensional coordinates of each reference feature point and marking the actual coordinates of each reference feature point;

[0077] In this embodiment, reference feature points with distinct geometric characteristics are selected at locations with stable internal structures of the tunnel boring machine, such as the cutterhead support frame and fixed beams inside the shield body. These reference feature points can be corner points, the center of circular holes, etc. High-precision measuring equipment is used to accurately measure and record the three-dimensional coordinates of these reference feature points.

[0078] Reference Figure 2 The calibration unit performs calibration and correction of spatial scan data based on reference feature points, specifically including:

[0079] Read the point cloud data composed of the spatial scan data; extract feature points from the point cloud data;

[0080] The extracted feature points are matched with the reference feature points to obtain the measured coordinates of each reference feature point in the point cloud data. This embodiment of the application extracts feature points from real-time scanning data using image recognition and point cloud processing algorithms. The extracted feature points are matched with pre-selected reference feature points, employing efficient matching algorithms such as the iterative nearest point algorithm to identify which specific feature point extracted from the point cloud corresponds to each reference feature point.

[0081] Based on the actual and measured coordinates of each reference feature point, the rotation matrix and translation vector of the point cloud data are calculated. In this embodiment, at least three reference feature points are preferred, and the rotation matrix and translation vector of the point cloud data relative to the reference coordinate system are calculated. The rotation matrix and translation vector describe the attitude deviation and position deviation of the point cloud data relative to the reference coordinate system, respectively.

[0082] The rotation matrix and translation vector are used to compensate and correct each spatial scan data in the point cloud data. Each spatial scan data is a three-dimensional coordinate; the three-dimensional coordinate is multiplied by the rotation matrix and then the translation vector is added to obtain the corrected spatial scan data. This application uses a feature selection unit and a calibration unit to process and calibrate the spatial scan data of real-time 3D scanning, ensuring the accuracy of the spatial scan data and avoiding the adverse effects of internal vibration interference of the tunnel boring machine or cutter changer robot or other environmental interference on the accuracy of the spatial scan data due to the positional displacement of the scanning components.

[0083] The path planning module performs arm span path planning for the tool-changing robot based on the spatial scanning data.

[0084] The path planning module includes an action decomposition unit, a path exploration unit, and a calculation unit;

[0085] The action decomposition unit is used to decompose the task of the tool changing robot into multiple sub-actions, and combine spatial scanning data to determine the target state and constraints of each sub-action.

[0086] The target state includes the starting coordinates and ending coordinates of the end effector; wherein, the starting coordinates are the coordinates of the end effector or a selectable range of coordinates when the sub-action begins execution; the ending coordinates are the coordinates of the end effector or a selectable range of coordinates when the sub-action is completed. The selection of both the starting and ending coordinates is based on a passable spatial range determined by spatial scan data; the constraints include spatial constraints, attitude constraints, and force constraints; wherein, spatial constraints are also implemented based on obstacles or interior walls of the space determined by spatial scan data.

[0087] Preferably, in this embodiment, the task of installing a new tool is broken down into the following sub-actions: tool gripping, transport posture adjustment, tool transport, installation posture calibration, and tool installation. Specifically, the spatial constraint for tool gripping is that the distance between the end effector and the obstacle is not less than 2cm; the posture constraint is that the tool's posture deviation before and after gripping is within ±3° to ensure stable gripping; the force constraint is that the clamping force is controlled between 18000-20000N to ensure the tool does not fall off (taking a tool mass of 150kg and a friction coefficient of 0.25 as an example, a clamping force of at least 18000N can ensure a safety factor of 3 times the minimum clamping force). The spatial constraint for transport posture adjustment is that the minimum distance between the tool and the arm span and surrounding obstacles is not less than 5cm to ensure transport safety; the posture constraint is that the tool's posture change range is within ±5° to meet subsequent installation requirements; the force constraint is that the force maintaining tool stability is controlled between 800-1200N to prevent tool wobbling (the force maintaining tool stability here is used to overcome inertia, centrifugal force, vibration interference, etc.). The spatial constraints for tool transport are as follows: the distance between the tool and the arm span and obstacles around the path is no less than 3cm; the attitude constraint is that the attitude fluctuation range of the tool during transport is within ±2°; the force constraint is that the driving force of the arm span is stable during transport, and the fluctuation range does not exceed 10% of the rated torque to ensure that the tool does not wobble. The spatial constraints for installation and orientation calibration are as follows: within a 5cm radius around the tool holder mounting hole, the arm span does not touch any obstacles; the attitude constraint is that the attitude deviation between the tool axis and the tool holder mounting hole axis is within ±1°, and the position deviation is within ±2mm; the force constraint is that the fine-tuning positioning force is controlled within 50-80N to achieve precise calibration. The spatial constraints for tool installation are as follows: during installation, the distance between the tool and obstacles around the tool holder is no less than 1cm; the attitude constraint is that the tool is inserted vertically into the tool holder, with a verticality deviation within ±1°; the force constraint is that the fixed torque error is within ±0.5%. This application breaks down the tool changing process in detail, dividing the task of the tool changing robot into multiple sub-actions. It fully considers the spatial, posture, and force requirements of each sub-action on the robot's arm movement, and adaptively sets target states and constraints for each sub-action by combining spatial scanning data. This achieves refined arm movement planning. Compared with traditional path planning schemes that only consider the overall movement path, this application can more accurately control the arm movement of the tool changing robot and reduce risks such as collisions and tool drops.

[0088] The path exploration unit explores alternative paths for each sub-action based on the target state and constraints of each sub-action, and generates a set of alternative paths for each sub-action; the set of alternative paths for any sub-action contains at least one alternative path for the corresponding sub-action.

[0089] The path exploration unit explores alternative paths for each sub-action based on the target state and constraints of each sub-action, specifically including:

[0090] Read the target state and obtain the start and end coordinates of the end effector of each sub-action;

[0091] Based on the path planning algorithm, considering the constraints, each sub-action plans at least M alternative paths; M is a positive integer; any alternative path for any sub-action is the movement of the end effector from the starting coordinate to the ending coordinate of the corresponding sub-action; in the embodiments of this application, the preferred path planning algorithm is any one of the A* algorithm, Dijkstra's algorithm, or Rapid Exploratory Random Tree (RRT) algorithm.

[0092] The computing unit generates an arm span path based on the candidate path set for each sub-action; specifically including:

[0093] The calculation unit selects one alternative path from the set of alternative paths for each sub-action, and combines all the selected alternative paths for the sub-actions into a pending arm span path.

[0094] Perform a stability test on any two adjacent sub-movements in the undetermined arm span path; if all two adjacent sub-movements in the undetermined arm span path pass the stability test, then mark the undetermined arm span path as an arm span path.

[0095] The stationarity detection is specifically as follows:

[0096] The acceleration change rate of each joint of the robot in two adjacent sub-actions is calculated. The calculation unit is configured with an acceleration change rate threshold. If the acceleration change rate of each joint is less than the acceleration change rate threshold, the two adjacent sub-actions pass the stability test. The acceleration change rate is the instantaneous change in acceleration of any joint when adjacent sub-actions are connected, divided by the time interval. For rotary joints, the acceleration of interest is angular acceleration; for translational joints, the acceleration of interest is linear acceleration. The acceleration change rate reflects the stability of the tool-changing robot's action switching. If the acceleration change rate of each joint is small, it indicates that there are no obvious abrupt changes in the robot's motion state, avoiding mechanical vibration caused by excessive acceleration or deceleration, and ensuring the continuity of the tool-changing action.

[0097] The computing unit generates no fewer than m reach paths and calculates the risk level of each reach path; the reach path with the lowest risk level is output as the reach path planning result for the tool-changing robot. This embodiment first uses 3D scanning to quickly obtain the space available for the reach mechanism of the tool-changing robot to extend; then, through path planning within the limited space, it controls the reach mechanism to drive the end effector to complete the tool-changing action along the optimal path, thus improving the efficiency of tool-changing robots applied in tunnel boring machines.

[0098] Reference Figure 4 The computing unit calculates the risk level of each arm span path, specifically including:

[0099] Calculate the lengths of each alternative path included in any arm span path and sum them to obtain the total arm span path length;

[0100] Based on the alternative paths for each sub-action in the arm span path, calculate the motion of each joint of the tool-changing robot when executing each sub-action; based on the motion of each joint, calculate the motion trajectory of each joint.

[0101] Read the coordinate range of each buffer and the core working area in the workspace; calculate the total number of times all joints pass through the buffer and the core working area based on the motion trajectory of each joint, and record the total number of times as the risk count;

[0102] Given the motion trajectory (i.e., arm span path) of the end effector, inverse kinematics can be used to calculate the motion sequence and amount of motion required for each joint to perform each sub-action along the arm span path. Based on the joint's motion amount, the range of coordinate changes of the joint in the workspace, i.e., the joint's motion trajectory, can be calculated. If the joint's motion trajectory intersects with a buffer zone or core working area, then the joint passes through the buffer zone or core working area.

[0103] The risk level of the arm span path is obtained by normalizing the total length of the arm span path and the number of risks, and then weighting and summing them.

[0104] This application considers both the total length of the arm span path and the number of risks to calculate the risk level for evaluating each arm span path. On the one hand, it encourages minimizing the total length of the arm span path to improve tool changing efficiency; on the other hand, it aims to minimize the joints of the tool changing robot traversing the buffer zone and core working area to reduce the risk of collisions.

[0105] 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.

[0106] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other modifications under the guidance of the present invention without departing from the spirit and scope of the present invention, and all of these modifications are within the scope of protection of the present invention.

Claims

1. A 3D scanning-based arm path planning system for a tool-changing robot within a confined space, characterized in that: It includes a perception module, a scanning module, and a path planning module; among which: The perception module is used to identify obstacle types; based on the obstacle type, the perception module divides the workspace of the tool-changing robot into different sub-regions; the types of the sub-regions include safe passage area, buffer zone, and core operation area; The scanning module acquires spatial scanning data for each sub-region based on 3D scanning; The sensing module is also used to select reference feature points; the scanning module calibrates the spatial scanning data based on the reference feature points; The path planning module performs arm span path planning for the tool-changing robot based on the spatial scan data, specifically including: The task of the tool-changing robot is broken down into multiple sub-actions; a set of alternative paths is generated for each sub-action. Generate arm span path based on the set of alternative paths for each sub-action; The risk level of each arm span path is calculated based on the number of times the joints of the tool-changing robot traverse different types of sub-regions, and the arm span path with the lowest risk level is selected as the planning result. The calculation of the risk level for each arm span path specifically includes: Calculate the lengths of each alternative path included in any arm span path and sum them to obtain the total arm span path length; Based on the alternative paths for each sub-action in the arm span path, calculate the motion of each joint of the tool-changing robot when executing each sub-action; based on the motion of each joint, calculate the motion trajectory of each joint. Read the coordinate range of each buffer and the core working area in the workspace; calculate the total number of times all joints pass through the buffer and the core working area based on the motion trajectory of each joint, and record the total number of times as the risk count; The risk level of the arm span path is obtained by normalizing the total length of the arm span path and the number of risks, and then weighting and summing them.

2. The 3D scanning-based arm span path planning system for a tool-changing robot in a limited space as described in claim 1, characterized in that: The sensing module includes a recognition unit; The identification unit is used to identify the types of obstacles in the tool-changing robot's workspace, specifically including: The system acquires images of obstacles in the workspace; the recognition unit is equipped with a trained target detection model, and the obstacle images are input into the trained target detection model to identify the obstacle types. The obstacle types include intrusion obstacles and structural obstacles; the intrusion obstacles include obstacles that accidentally intrude into the workspace of the tool-changing robot; the structural obstacles include the internal structure of the robot's work cabin.

3. The 3D scanning-based arm span path planning system for a tool-changing robot in a limited space as described in claim 2, characterized in that: The perception module further includes an evaluation unit; the evaluation unit is used to divide the workspace of the tool-changing robot into different sub-regions, specifically including: The coordinate range of each obstacle is obtained; the recognition unit is also equipped with a binocular vision device; the recognition unit calculates the coordinate range of each obstacle based on the binocular vision device and transmits it to the evaluation unit; Based on the coordinate range and obstacle type of each obstacle, the workspace is divided into different types of sub-regions, and the coordinate range of each sub-region is recorded; specifically including: The core working area is divided with the cutterhead of the tunnel boring machine as the center; outside the core working area, a buffer zone is divided with any structural obstacle or intrusion obstacle as the center; in the work space, the area outside the buffer zone and the core working area is the safe passage zone.

4. The 3D scanning-based arm span path planning system for a tool-changing robot in a confined space as described in claim 3, characterized in that: The scanning module includes a scanning component unit; the scanning component unit is used to perform 3D scanning on each sub-region to obtain spatial scanning data of each sub-region, specifically including: Set the scan resolution for each type of sub-region; read the type and coordinate range of each sub-region; The workspace of the tool-changing robot is 3D scanned, and the coordinates of the scanned points are detected in real time. The type of the sub-region where the scan point is located is identified based on the coordinates of the scan point, and the scan resolution is dynamically adjusted based on the type of the sub-region.

5. The 3D scanning-based arm span path planning system for a tool-changing robot in a limited space as described in claim 4, characterized in that: The sensing module further includes a feature selection unit; the scanning module further includes a calibration unit; The feature selection unit is used to select reference feature points in the workspace of the tool-changing robot, specifically including: selecting no less than n reference feature points before the 3D scan begins; n is a positive integer; measuring the three-dimensional coordinates of each reference feature point and marking the actual coordinates of each reference feature point; The calibration unit performs calibration and correction of spatial scan data based on reference feature points, specifically including: Read the point cloud data composed of the spatial scan data; extract feature points from the point cloud data; The extracted feature points are matched with the reference feature points to obtain the measured coordinates of each reference feature point in the point cloud data; Based on the actual and measured coordinates of each reference feature point, calculate the rotation matrix and translation vector of the point cloud data; The rotation matrix and translation vector are used to compensate and correct each spatial scan data in the point cloud data.

6. The 3D scanning-based arm span path planning system for a tool-changing robot in a limited space as described in claim 5, characterized in that: The path planning module includes an action decomposition unit; the action decomposition unit is used to decompose the task of the tool changing robot into multiple sub-actions, and combine spatial scanning data to determine the target state and constraints of each sub-action; The target state includes the starting coordinates and ending coordinates of the end effector; the constraints include spatial constraints, attitude constraints, and force constraints.

7. The 3D scanning-based arm span path planning system for a tool-changing robot in a confined space as described in claim 6, characterized in that: The path planning module further includes a path exploration unit; the path exploration unit explores alternative paths for each sub-action based on the target state and constraints of each sub-action, and generates a set of alternative paths for each sub-action; the set of alternative paths for any sub-action contains at least one alternative path for the corresponding sub-action. The path exploration unit explores alternative paths for each sub-action based on the target state and constraints of each sub-action, specifically including: Read the target state and obtain the start and end coordinates of the end effector of each sub-action; Based on the path planning algorithm, considering the constraints, each sub-action plans at least M alternative paths; M is a positive integer; any alternative path for any sub-action is the movement of the end effector from the starting coordinate to the ending coordinate of the corresponding sub-action.

8. The 3D scanning-based arm span path planning system for a tool-changing robot in a confined space as described in claim 7, characterized in that: The path planning module further includes a calculation unit; the calculation unit generates an arm span path based on the candidate path set for each sub-action; specifically including: The calculation unit selects one alternative path from the set of alternative paths for each sub-action, and combines all the selected alternative paths for the sub-actions into a pending arm span path. Perform a stability test on any two adjacent sub-movements in the undetermined arm span path; if all two adjacent sub-movements in the undetermined arm span path pass the stability test, then mark the undetermined arm span path as an arm span path. The computing unit generates no fewer than m arm span paths and calculates the risk level of each arm span path; the arm span path with the lowest risk level is output as the arm span path planning result of the tool changing robot.

9. The 3D scanning-based arm span path planning system for a tool-changing robot in a confined space as described in claim 8, characterized in that: The stationarity detection is specifically as follows: Calculate the rate of change of acceleration of each joint of the robot in two adjacent sub-actions; the calculation unit is configured with an acceleration rate of change threshold; if the rate of change of acceleration of each joint is less than the acceleration rate of change threshold, then the two adjacent sub-actions pass the stability test.

Images