Mobile mechanical arm pressing plate throwing and withdrawing operation method and system based on segmented operation

Abstract

The invention discloses a moving mechanical arm pressing plate throwing and withdrawing operation method and system based on segmented operation, space positions of pressing plate binding post locking nuts and connecting pieces are obtained through visual recognition, a coordinate transformation module converts pixel coordinates into three-dimensional positions and postures under a mechanical arm base coordinate system, position control is adopted by a mechanical arm in the approaching stage, and the three-dimensional positions and postures of the mechanical arm in the approaching stage are controlled. Force control is switched in the contact stage, and segmented automatic operation of locking nut unscrewing, connecting piece pulling-out, connecting piece pulling-in and locking nut screwing is achieved according to a force threshold value, torque changes and displacement changes. And the task instruction module is used for receiving a pressing plate'input 'or'quit' operation instruction, so that the system can independently execute a pressing plate input process or quit process according to the instruction, and the operation flexibility is improved. And the visual verification module compares and confirms the states before and after the operation to ensure the accuracy of the commissioning and retreating operation. According to the invention, single-arm segmentation operation in the whole process of inputting and withdrawing the pressing plate is realized, and the device has high safety and reliability and is suitable for electric power inspection and industrial operation scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of robot intelligent control and automatic operation, and in particular to a mobile manipulator pressboard putting-off operation method and system based on segmented operation, which is used for automatic putting-off operation of pressboard of power control cabinet and power distribution system. BACKGROUND

[0002] In the process of power system operation and maintenance, pressboard putting-off is a key step to ensure the safety of equipment and function switching. The traditional pressboard operation is generally completed by manual operation, and the operator needs to complete the operations of loosening the lock nut of the pressboard binding post, moving the connecting piece, closing the switch and tightening, etc. However, manual operation has obvious shortcomings: firstly, the manual operation force is uncontrollable, which may cause the lock nut of the pressboard binding post to be worn or the connecting piece to be deformed; secondly, manual intervention in live or complex environments has the risk of electric shock and misoperation; thirdly, long-term manual operation is repetitive and low in efficiency, which is difficult to meet the needs of modern intelligent substation and unmanned operation and maintenance.

[0003] Although there are attempts to use a manipulator to replace manual operation of the pressboard in existing research, there are the following technical bottlenecks: (1) most of them are double-armed or have complex end structures, which have poor space adaptability; (2) the action control uses position or force as a single control variable, and lacks a phased hierarchical strategy; (3) there is a lack of self-adaptive control logic in the process of action, and the success rate is low when facing different fastening degrees or structural differences; (4) most existing systems use a fixed sequence of putting-off process, and lack the response ability to external task instructions, which cannot execute the "put-in" or "exit" operation according to the equipment operating state, and do not have the instruction control characteristics required by the dispatching system; (5) in terms of visual recognition and tactile control, most schemes lack a clear coordinate conversion model and threshold determination mechanism, making it difficult to achieve stable and reliable segmented force control execution.

[0004] Therefore, it is necessary to develop a mobile manipulator pressboard putting-off method and system based on visual perception, coordinate mapping and segmented force control execution, so that the manipulator can complete the whole process of loosening, moving and tightening under the condition of single arm, and determine the result through visual verification, thereby realizing high safety, universality and flexible control. SUMMARY

[0005] The present application provides a mobile manipulator pressboard putting-off operation method and system based on segmented operation, which realizes automatic identification, accurate approach, segmented loosening / tightening and moving-in / moving-out operation of the pressboard binding post lock nut and connecting piece through the fusion of visual perception, coordinate mapping, task instruction analysis, planning control and force feedback, thereby safely and reliably completing the exit and entry process of the pressboard.

[0006] Technical solution: This invention proposes a mobile robotic arm pressure plate deployment and retraction operating system based on segmented operation, including a six-degree-of-freedom mobile robotic arm, a vision perception module, a coordinate transformation module, an end effector module, a force perception module, a segmented deployment and retraction control module, a task instruction module, and a vision verification module.

[0007] The six-degree-of-freedom mobile robotic arm is used to perform the physical operation of throwing and retracting the pressure plate, realizing multi-stage single-arm movements;

[0008] The visual perception module is used to acquire images of the target area and identify the positions of the pressure plate terminal block lock nut and connecting piece;

[0009] The coordinate transformation module is used to convert visual coordinates into robotic arm base coordinates and provide target pose for the motion planning of the six-degree-of-freedom mobile robotic arm.

[0010] The end operation module is used to loosen or tighten the terminal lock nut during the process of the pressure plate being removed and engaged, and to push out or push in the pressure plate connecting piece.

[0011] The force sensing module is used to detect contact force, lateral force and torque information during the approach, contact and operation of the six-degree-of-freedom mobile robotic arm, and to provide feedback information for segmented deployment and retraction control;

[0012] The segmented deployment and retraction control module is used to divide the deployment and retraction process of the pressure plate into stages of approach, contact, loosening, pulling out, pulling in, tightening and retraction based on visual and force perception information, and to switch between position control and force control.

[0013] The task instruction module is used to receive the pressure plate "put in" or "exit" operation instruction and provide the instruction to the segmented input and output control module, and select to execute the pressure plate input process or the pressure plate exit process according to different instructions.

[0014] The visual verification module is used to acquire images after the operation is completed, and to determine whether the pressure plate is in the correct state by comparing the images before and after the operation.

[0015] Furthermore, the physical operation of the pressure plate engagement and disengagement includes a pressure plate disengagement operation and a pressure plate engagement operation; the pressure plate disengagement operation includes the stage of loosening the terminal lock nut and the stage of pulling out the connecting piece, and the pressure plate engagement operation includes the stage of pulling the connecting piece into the closed position and the stage of tightening the terminal lock nut.

[0016] Furthermore, the visual perception module uses a depth camera to simultaneously acquire color and depth images, and performs four steps: image preprocessing, target recognition, pixel coordinate extraction, and three-dimensional coordinate back projection, as detailed below:

[0017] A depth camera acquires an RGB-D image set of the pressure plate area. The RGB images are used for shape recognition of the lock nuts and connecting pieces. The pixel coordinates of the lock nuts and connecting pieces of the pressure plate terminal block are extracted by a visual algorithm. The depth images are used to obtain their depth values.

[0018] Using a pinhole model, pixel coordinates are back-projected onto the three-dimensional space of the camera coordinate system. The three-dimensional points obtained based on the depth values ​​represent the actual spatial positions of the lock nuts or connecting pieces in the camera coordinate system.

[0019] The three-dimensional coordinates obtained by back projection and the identified target categories are output to the coordinate transformation module to provide data input for subsequent coordinate transformation, trajectory planning and robotic arm control. The target categories include lock nuts or connecting pieces.

[0020] Furthermore, in order to obtain the target pose, the coordinate transformation module uses the directional features of the locking nut or connecting pieces identified by the visual perception module, combined with the camera attitude rotation matrix. After transforming the direction vector, the target attitude matrix is ​​obtained from... Received, among which The orientation matrix of the target in the camera coordinate system is defined as the direction feature of the lock nut or the connecting piece, which is the direction of the lock nut normal vector or the direction of the connecting piece edge.

[0021] Furthermore, the specific process of the force sensing module is as follows:

[0022] Force sensors detect contact force, lateral force, and torque signals in real time. ,in It is a triaxial contact force. It is a triaxial torque;

[0023] The original signal is biased and proportionally calibrated using a sensor calibration matrix to ensure that the measured force conforms to a linear model. , To calibrate the gain matrix, It is a zero-biased vector;

[0024] Transform the force sensor coordinate system to the end-effector coordinate system or the robot arm base coordinate system;

[0025] Time-domain filtering is performed on the force / torque signal to eliminate high-frequency noise and obtain a smooth force signal;

[0026] Based on the filtered force and torque signals, the segmented engagement and disengagement control module sets multiple judgment thresholds to distinguish between the loosening, pulling out, pulling in, and tightening stages.

[0027] Furthermore, the segmented deployment and retraction control module performs segmented control of the six-degree-of-freedom mobile robotic arm's motion based on visual positioning, task instructions from the task instruction module, and force feedback from the force sensing module. This is divided into an approach phase and a contact and operation phase, as detailed below:

[0028] During the approach phase, an approach path is planned based on the three-dimensional position and target pose provided by the coordinate transformation module, so that the end effector module of the six-degree-of-freedom mobile robotic arm gradually approaches the lock nut or connecting piece, using position control.

[0029] During the contact and operation phase, once the end effector module enters the contact area, the force sensing module begins to detect the contact force, lateral force, and torque in real time. When the contact force reaches the set threshold, the module switches from position control to force control. The force control adopts an impedance control model. The entire force sensing process runs in real time within the control cycle, providing stage judgment signals for the segmented deployment and retraction control module. When an abnormal force peak or torque over-limit is detected, the protection mechanism is immediately triggered, causing the six-degree-of-freedom mobile robotic arm to perform a retraction action.

[0030] Furthermore, the segmented deployment and deactivation control module sets multi-level judgment thresholds to distinguish between the loosening, pulling out, pulling in, and tightening stages, specifically:

[0031] During the lock nut loosening stage, the first contact force threshold is used. Contact establishment is determined, and the loosening torque criterion is used simultaneously. ,in The relaxation threshold, It is obtained by calculating the difference between continuous torque signals;

[0032] During the continuous strip release phase, a second contact force threshold is adopted. Monitoring the displacement changes in the direction of continuous pull-out and reaction force The allocation was completed to meet the joint criteria. ,in Displacement threshold The reaction force threshold;

[0033] During the continuous insertion stage, the trend of the insertion force with displacement is monitored to determine whether the continuous piece has entered the closing position. When the force change tends to stabilize and the displacement reaches the set position, the insertion is determined to be completed.

[0034] During the tightening phase of the lock nut, a third contact force threshold is used. Establish contact and employ the torque convergence criterion. This is used to determine whether the lock nut is tightened completely. To tighten the threshold.

[0035] Furthermore, the task instruction module parses the input "input / output" command into a Cmd signal, i.e., Cmd=Strap_Enable or Cmd=Strap_Disable. Then, the Cmd signal is passed to the segmented input / output control module, which decides whether to perform a pressure plate output operation or a pressure plate input operation.

[0036] This invention also discloses a method for the deployment and retraction of a pressure plate by a mobile robotic arm based on segmented operation, comprising the following steps:

[0037] Step S1: Acquire images of the pressure plate area and identify the pixel positions of the pressure plate terminal blocks, lock nuts, and connecting pieces;

[0038] Step S2: Convert the pixel position into robotic arm base coordinates;

[0039] Step S3: The task instruction module parses the instruction to generate either the pressure plate input instruction Cmd=Strap_Enable or the pressure plate exit instruction Cmd=Strap_Disable;

[0040] Step S4: Based on the generated instruction Cmd, select either exit operation or engage operation and plan the approach trajectory of the six-degree-of-freedom mobile robotic arm;

[0041] Step S5: Detect the contact state through the force sensing module, and switch to force control when the contact force reaches the set threshold;

[0042] Step S6: When Cmd=Strap_Disable, the lock nut is loosened and the connecting piece is pulled out in stages;

[0043] Step S7: When Cmd=Strap_Enable, complete the segmented insertion of the connecting piece and tightening of the lock nut;

[0044] Step S8: After the operation is completed, the visual verification module is used to determine the pressure plate's engagement / disengagement status;

[0045] Step S9: If the verification fails, replan the trajectory of the six-DOF mobile robotic arm and repeat steps S3 to S7.

[0046] Compared with the prior art, the present invention has the following technical effects:

[0047] 1. This invention utilizes visual positioning to achieve precise identification and spatial positioning of the pressure plate lock nut and connecting piece. It employs multi-level force thresholds to segment and protect key actions such as loosening the lock nut, extending and retracting the connecting piece, and tightening the lock nut. Furthermore, it uses Cmd commands to drive switchable engagement and disengagement processes, achieving automation and controllability throughout the entire pressure plate engagement and disengagement process. This invention improves the reliability and efficiency of pressure plate engagement and disengagement operations while ensuring the safety of electrical equipment, making it suitable for unmanned operation and maintenance scenarios in power control cabinets and power distribution systems.

[0048] 2. This invention adopts a single-arm structure to realize multi-step operation, which is compact and occupies little space; segmented control combined with force feedback realizes smooth and safe operation switching; visual recognition and coordinate mapping are combined to improve operation accuracy and positioning reliability; visual verification realizes non-contact result confirmation, avoids electrical interference, and improves versatility; the control logic is portable and can be quickly deployed on various types of devices, which has good application prospects. Attached Figure Description

[0049] Figure 1 This is a block diagram of the overall structure of the present invention.

[0050] Figure 2 This is a flowchart of the steps of the present invention.

[0051] Figure 3 This is a schematic diagram of the coordinate mapping based on visual information according to the present invention;

[0052] Figure 4 This is a physical image of an embodiment of the present invention. Detailed Implementation

[0053] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0054] This invention discloses a method and system for the deployment and retraction of a pressure plate by a mobile robotic arm based on segmented operation. See [link to relevant documentation]. Figures 1-4 It mainly includes a six-degree-of-freedom mobile robotic arm, a vision perception module, a coordinate transformation module, an end effector module, a force perception module, a segmented deployment and retraction control module, a task instruction module, and a vision verification module.

[0055] The six-degree-of-freedom (DOF) mobile robotic arm is used to execute motion paths; the vision perception module is used to detect the position of the pressure plate terminal lock nut and connecting piece; the coordinate transformation module is used to convert the pixel coordinates and depth information obtained by visual recognition into the robotic arm's base coordinates and provide the target pose for the motion planning of the six-DOF mobile robotic arm; the force perception module is used to obtain contact force, lateral force and torque information; the segmented deployment and retraction control module is used to coordinate the segmented motion according to the position and force feedback; the task instruction module is used to select the exit process or deployment process according to the input command; the vision verification module is used to compare the state before and after the pressure plate deployment and retraction.

[0056] The workflow of this invention includes steps such as visual positioning, coordinate mapping, task instruction parsing, robotic arm approach, contact detection, exit / entry operation, and visual verification. First, the visual perception module acquires RGB-D images of the pressure plate area, identifies the pixel positions of the locking nut and connecting piece using a target detection algorithm, and obtains the depth value of that position from the depth map. After preprocessing, filtering, and multi-frame fusion, the visual perception module outputs stable pixel coordinates and depth data for spatial positioning. Subsequently, the coordinate transformation module converts the pixel coordinates and depth values ​​into three-dimensional points in the camera coordinate system, and then, based on the pre-calibrated extrinsic parameter relationship between the camera and the robotic arm, transforms these three-dimensional points into a three-dimensional position in the robotic arm's base coordinate system. Simultaneously, the coordinate transformation module converts the target posture based on the identified object orientation (such as the locking nut normal direction or the connecting piece edge direction), enabling the robotic arm to not only approach the correct position but also maintain the correct orientation. The task instruction module receives "entry / exit" instructions from external sources and parses them into Cmd control signals. The value of Cmd determines the selection of subsequent robotic arm action flows, where Exit corresponds to the exit action and Enter corresponds to the entry action. During the approach phase, the segmented deployment and retraction control module plans an approach path based on the three-dimensional position and attitude provided by the coordinate transformation module, enabling the robotic arm's end effector module to gradually approach the locking nut or connecting piece. In this phase, the robotic arm employs position control, ensuring a safe and stable approach through trajectory smoothing and speed limiting. Once the end effector module enters the contact area, the force sensing module begins real-time detection of contact force, lateral force, and torque. When the system detects that the contact force has reached a set threshold, it switches from position control to force control, ensuring the robotic arm remains compliant after contact with the structure, preventing damage to the pressure plate and connecting piece.

[0057] When Cmd = Strap_Disable, the system enters the exit process, which includes a loosening phase and a pulling phase. In the loosening phase, after the end effector module establishes stable contact with the lock nut, it applies a reverse rotation. The force sensing module continuously monitors the torque trend; when the torque continuously decreases and reaches a stable low value, it determines that the lock nut has been loosened. Simultaneously, the system also monitors for abnormal torque spikes or exceeding safety thresholds. If a dangerous trend is detected, the robotic arm will immediately retract to prevent the lock nut from stripping or structural damage. After loosening, the robotic arm moves to the side of the connecting piece and enters the pulling phase. The end effector module applies a lateral pulling force, and the segmented deployment / retraction control module simultaneously monitors the pulling-out displacement and reaction force of the connecting piece. When the displacement of the connecting piece exceeds a set value and the reaction force decreases to a safe range, it determines that the connecting piece has disengaged from the contact, and the exit operation is complete.

[0058] When Cmd = Strap_Enable, the system enters the deployment process, which includes the connecting piece insertion stage and the lock nut tightening stage. During the connecting piece insertion stage, the robotic arm pushes the connecting piece into the corresponding contact position along the planned path. The system monitors the insertion depth and reaction force changes of the connecting piece in real time. When the connecting piece reaches the expected position and the contact force change tends to stabilize, the insertion is considered complete. During the lock nut tightening stage, the end effector module contacts the lock nut and applies a forward tightening action. The force sensing module monitors the torque increase; when the torque reaches the preset tightening value and enters a stable phase, the lock nut tightening is considered complete. If the torque continues to rise abnormally or exceeds the limit, the system will immediately stop and retract to avoid structural damage.

[0059] After the entire exit or engagement operation is completed, the system enters the visual verification phase. The visual verification module acquires images after the operation and compares them with images before the operation. It determines whether the action was successful by using indicators such as changes in the position of the connecting pieces, differences in the lock nut's posture, and the amount of structural change. If verification fails, the system will replan the trajectory and re-execute from the approach phase to improve overall reliability.

[0060] The embodiments of the present invention will be described in detail below:

[0061] 1. A six-degree-of-freedom (DOF) robotic arm, serving as the execution body, has six independent motion joints, enabling precise operation in any posture within three-dimensional space. The end effector module at the end of the robotic arm can perform actions such as approaching, loosening, and tightening the locking nuts of the pressure plate terminals, as well as pushing out and pushing in the pressure plate connecting plates, according to task requirements. This invention uses the robotic arm's base coordinate system as a unified reference coordinate system. The robotic arm operates based on the position of the target point in the base coordinate system obtained from the coordinate transformation module. Joint angles are solved by inverse kinematics This enables the end effector to move accurately to the target position obtained by visual recognition and coordinate transformation. This represents the inverse kinematics solution function, used to calculate the corresponding joint angle solutions based on the desired pose of the robotic arm's end effector in space. This represents the desired attitude matrix of the target end effector in the robot arm's base coordinate system. It describes the spatial orientation of the end effector module relative to the base coordinate system and is typically a 3×3 rotation matrix. , It represents the Special Orthogonal Group in 3D, which describes the rotation of rigid bodies in three-dimensional space.

[0062] 2. The visual perception module, installed near the end effector of the six-DOF mobile robotic arm, uses a depth camera to simultaneously acquire color and depth images. The visual perception process includes four steps: image preprocessing, target recognition, pixel coordinate extraction, and 3D coordinate back projection, as detailed below:

[0063] First, a depth camera acquires an RGB-D image set of the pressure plate area. The RGB images are used for shape recognition of the lock nuts and connecting pieces, while the depth images are used to obtain their depth values. After preprocessing such as filtering, brightness equalization, and edge enhancement, the pixel coordinates of the pressure plate terminal lock nuts and connecting pieces are extracted using visual algorithms. To ensure robustness of recognition, the system fuses recognition results from multiple consecutive frames to obtain the final pixel position and depth values. .

[0064] Subsequently, using a pinhole model, the pixel coordinates are back-projected onto the camera coordinate system in three-dimensional space. The camera intrinsic parameter matrix is ​​as follows:

[0065]

[0066] in, For camera focal length, The coordinates of the main point.

[0067] Based on depth value The camera coordinates of the 3D points are calculated as follows:

[0068]

[0069] The obtained three-dimensional points This indicates the actual spatial position of the lock nut or connecting piece in the camera coordinate system.

[0070] Finally, the visual perception module outputs the three-dimensional coordinates obtained by back projection and the identified target category (lock nut or connecting piece) to the coordinate transformation module, providing data input for subsequent coordinate transformation, trajectory planning and robotic arm control.

[0071] 3. The coordinate transformation module converts the camera coordinates output by the vision perception module into three-dimensional spatial positions in the robot arm's base coordinate system. The coordinate transformation process includes four steps: extrinsic parameter calibration, coordinate system definition, homogeneous transformation, and attitude recovery.

[0072] First, during system initialization, the rigid transformation relationship between the depth camera coordinate system and the robot arm base coordinate system is obtained through extrinsic parameter calibration. The calibration result is obtained from the rotation matrix. With translation vector Composed of, and forming a homogeneous extrinsic parameter matrix:

[0073]

[0074] in This represents the three-dimensional rotation matrix of the camera coordinate system relative to the robot arm base coordinate system. The rotation matrix satisfies the orthogonality condition and has a determinant of 1, and is used to describe the pure rotation relationship between the two coordinate systems. This indicates the position of the camera coordinate system origin in the robot arm's base coordinate system.

[0075] The 3D point representation of the camera coordinate system output by the visual perception module is as follows:

[0076]

[0077] The coordinate transformation module extends it to homogeneous coordinate form:

[0078]

[0079] Subsequently, a coordinate system mapping is achieved using a homogeneous transformation. The three-dimensional coordinates in the robot arm's base coordinate system can be calculated using the following formula: And thus obtain ,in This indicates the position of the target point in the robot arm's base coordinate system, and is the final input position for robot arm planning.

[0080] To further obtain the target pose, this invention utilizes the locking nut or connecting piece orientation features (e.g., normal vectors or edge directions) identified by the visual perception module, combined with the camera pose rotation matrix. After transforming the direction vector, the target attitude matrix is ​​obtained from... Received, among which This is the orientation matrix of the target in the camera coordinate system. This coordinate-attitude joint mapping ensures that the end effector module of the robotic arm has the correct spatial orientation when approaching the locking nut or connecting piece.

[0081] The coordinate transformation module ultimately outputs the three-dimensional position and orientation of the target point in the robot arm's base coordinate system, enabling the robot arm trajectory planning to run based on a unified robot arm base coordinate system, thereby ensuring the accuracy and continuity of the operation path.

[0082] 4. Force sensing module, installed on the end effector module of the robotic arm, is used to detect changes in contact force and torque in real time during the contact process between the robotic arm and the locking nut or connecting piece. The entire force sensing process includes five steps: six-axis force / torque acquisition, coordinate system transformation, filtering, threshold judgment, and trend analysis.

[0083] First, the force sensor detects contact force, lateral force, and torque signals in real time. ,in It is a triaxial contact force. This is a three-axis torque. To obtain the effective quantity relevant to the operation, the raw signal is offset compensated and proportionally calibrated using a sensor calibration matrix to ensure that the measured force satisfies a linear model. ,in To calibrate the gain matrix, It is a zero-biased vector.

[0084] Subsequently, to obtain the force in a specific direction of interest, the force sensor coordinate system is transformed to the coordinate system of the robotic arm's end effector or the robotic arm's base coordinate system, ensuring that the force determination process is performed in a unified spatial coordinate system. The transformation relationship is as follows: ,in Force in the sensor coordinate system Let be the rotation matrix of the tool coordinate system.

[0085] During real-time monitoring, this invention performs time-domain filtering on the force / torque signal to eliminate high-frequency noise and obtain a smooth force signal. The filtering method employs a first-order low-pass filter: ,in The filtered force These are the filter coefficients.

[0086] 5. The task instruction module is used to parse the input "input / output" command into a Cmd signal, i.e., Cmd=Strap_Enable or Cmd=Strap_Disable. Then, the Cmd signal is passed to the segmented input / output control module, which decides whether to execute the pressure plate output process or the pressure plate input process. This makes the pressure plate input / output process no longer a fixed sequence process, but is executed as needed according to the operation and maintenance schedule.

[0087] 6. Segmented deployment and retraction control module: Based on visual positioning, task command Cmd, and force feedback, the robot arm's movement is controlled in segments. Position control is used during the approach phase, and force control is used during the contact and manipulation phases. Action switching is achieved through a finite state machine. Cmd is used to select whether to enter the "exit process" or the "entry process". The segmented entry and exit control module combines the contact force and torque information fed back by the force sensing module, and determines whether each segment operation is completed based on preset thresholds and changing trends, and performs protection retraction or re-approach if necessary. This invention uses a contact judgment formula:

[0088] Position control

[0089] Force control

[0090] in The corresponding contact thresholds for each stage are set (as detailed below) to enable automatic switching of force control mode.

[0091] Based on the filtered force and torque signals, this invention sets multi-level judgment thresholds to distinguish between the loosening, pulling out, pulling in and tightening stages.

[0092] During the lock nut loosening stage, the present invention employs a first contact force threshold. Contact establishment is determined, and the loosening torque criterion is used simultaneously. ,in The relaxation threshold, It is obtained by calculating the difference between continuous torque signals.

[0093] During the continuous strip release phase, a second contact force threshold is adopted. This invention monitors the displacement changes in the direction of continuous pull-out. and reaction force The allocation was completed, satisfying the joint criteria. ,in Displacement threshold This is the reaction force threshold.

[0094] During the continuous insertion stage, this invention monitors the changing trend of insertion force with displacement to determine whether the continuous piece has entered the closing position. When the force change tends to stabilize and the displacement reaches the set position, the insertion is determined to be complete.

[0095] During the tightening stage of the lock nut, this invention employs a contact force threshold. Establish contact and employ the torque convergence criterion. This is used to determine whether the lock nut is tightened completely. To tighten the threshold.

[0096] This invention employs an impedance control model in the force control stage. To ensure the robotic arm remains compliant and stable during contact, among other things, This represents the desired output control force (or equivalent force) at the end of the robotic arm, which is the target force calculated by the impedance controller based on the current position and desired state, used to drive the robotic arm to interact with the external environment. This represents the stiffness coefficient (proportional gain), used to describe the response strength of the robotic arm's end effector to displacement deviations. It represents the damping coefficient (differential gain), which is used to suppress oscillations during end motion and improve the stability and smoothness of the contact process. This indicates the desired position (reference displacement), which is the target position that the end-effector module hopes to achieve in the current segment operation. Indicates the actual position of the robotic arm's end effector. Indicates the desired speed. This indicates the actual speed at the end of the robotic arm.

[0097] The entire force sensing process operates in real time within the control cycle, providing stage judgment signals for the segmented deployment and retraction control module. When an abnormal force peak or torque over-limit is detected, the invention immediately triggers a protection mechanism, causing the robotic arm to perform a retraction action to ensure the safety of the pressure plate structure and the system.

[0098] 7. End-effector module, installed at the end of the six-degree-of-freedom mobile robotic arm, is used to directly contact the terminal lock nut and connecting piece. It applies axial force, lateral force and rotational force through clamping, prying and other methods to complete specific operations such as loosening / tightening the lock nut and prying out / in the connecting piece.

[0099] 8. A visual verification module is used to re-capture images of the pressure plate area after the pressure plate is engaged or disengaged, and compare them with the images before the operation. It determines whether the pressure plate state has changed from engaged to disengaged or vice versa by observing changes in the position of the connecting pieces, the attitude of the lock nut, and the surrounding structure. This serves as the final criterion for determining the success of the operation. The change is identified using... ,when The operation is considered successful at that time.

[0100] This invention provides a method for the deployment and retraction of a pressure plate using a mobile robotic arm based on segmented operations, specifically including:

[0101] Step S1: Acquire images of the pressure plate area and identify the pixel positions of the pressure plate terminal block lock nut and connecting piece.

[0102] Step S2: Convert the pixel position into robotic arm base coordinates.

[0103] Step S3: The task instruction module parses the instruction and generates a pressure plate input (Cmd=Strap_Enable) or pressure plate exit instruction (Cmd=Strap_Disable).

[0104] Step S4: Based on the generated instructions, select either exit or engage the operation and plan the robotic arm's approach trajectory;

[0105] Step S5: Detect the contact state through the force sensing module, and switch to force control when the contact force reaches the set threshold.

[0106] Step S6: When Cmd=Strap_Disable (exit command), the lock nut is loosened and the connecting piece is pulled out in stages.

[0107] Step S61: During the lock nut loosening stage, when the end operation module contacts the lock nut and the force sensing module detects that the contact force reaches the first contact force threshold... Upon confirming reliable contact between the end effector and the lock nut, the control mode switches from position control to force control. The segmented engagement / disengagement control module controls the end effector to apply the loosening torque, while the force sensing module monitors torque changes in real time during rotation. The control module determines that the lock nut has been loosened; if the torque changes abnormally or exceeds the safety limit, the control module immediately stops loosening and performs a retraction action.

[0108] Step S62: During the continuous strip release stage, when the end operation module moves to the side of the continuous strip and detects that the lateral contact force reaches the second contact force threshold... At this time, the system enters the pull-out operation. The segmented deployment and retraction control module controls the end effector module to apply a lateral pull-out force along the continuous pull-out direction, and the force sensing module and the robotic arm position feedback module jointly monitor the displacement in the continuous pull-out direction. and reaction force According to the changes Once it is determined that the connecting piece has completely disengaged from the closed position, the connecting piece removal stage in the pressure plate exit operation is completed.

[0109] Step S7: When Cmd=Strap_Enable (engagement command), the continuous strip is engaged and the lock nut is tightened in stages;

[0110] Step S71: During the continuous insertion phase, the robotic arm drives the end-effector to move along the continuous closing direction according to the planned path, so that the continuous pieces are gradually inserted into the corresponding contact positions. By monitoring the displacement of the continuous pieces and the changes in contact reaction force, it is determined whether the insertion force is within a reasonable range. When the position of the continuous pieces reaches the predetermined closing position and the change in contact force tends to stabilize, the continuous insertion is considered complete.

[0111] Step S72: During the lock nut tightening stage, when the end operation module contacts the lock nut and the force sensing module detects that the contact force reaches the third contact force threshold... At this time, the system enters the tightening operation. The segmented engagement / disengagement control module controls the end-operation module to apply positive tightening torque, and the force sensing module monitors the torque increase during the tightening process in real time, and adjusts accordingly. The lock nut is determined to be tightened. If the torque continues to rise abnormally or exceeds the safety limit, the control module immediately stops tightening and performs retraction to avoid damaging the lock nut and pressure plate structure.

[0112] Step S8: After the operation is completed, the visual verification module is used to determine the pressure plate engagement / disengagement status.

[0113] Step S9: If the verification fails, replan the robotic arm trajectory and repeat steps S4 to S8.

[0114] like Figure 4 As shown, the present invention has achieved stable and reliable identification, exit and entry operations in a real power cabinet environment, verifying the feasibility and practicality of the system.

[0115] This invention has been experimentally verified to ensure that the robotic arm can stably complete the plate removal and insertion processes under different plate structures, tightness levels, and friction conditions. The entire operation is smooth and precisely controlled. Multi-force threshold judgment and segmented control effectively avoid fault scenarios such as plate jamming and lock nut stripping. This invention can not only be used for plate operation in power control cabinets, but also be extended to automated tasks such as switch mechanism levers, rotary latches, and pluggable components, demonstrating broad application value in intelligent operation and maintenance scenarios.

[0116] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent transformations or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A mobile robotic arm pressure plate deployment and retraction operating system based on segmented operation, characterized in that, The six-degree-of-freedom mobile manipulator, the visual perception module, the coordinate transformation module, the end operation module, the force perception module, the segmented control module, the task instruction module and the visual verification module are comprised. The six-degree-of-freedom mobile manipulator is used for executing the physical operation of the pressing plate and realizing the single-arm multi-stage action. The visual perception module is used for collecting the image of the target area and identifying the position of the pressing plate terminal post lock and the connecting piece. The coordinate transformation module is used for converting the visual coordinates into the base coordinates of the manipulator and providing the target pose for the motion planning of the six-degree-of-freedom mobile manipulator. The end operation module is used for executing the loosening or tightening operation of the terminal post lock and the pushing-out or pushing-in operation of the pressing plate connecting piece during the pressing plate withdrawal and input process. The force perception module is used for detecting the contact force, lateral force and torque information change during the approach, contact and operation of the six-degree-of-freedom mobile manipulator and providing the feedback information for the segmented control. The segmented control module is used for dividing the pressing plate withdrawal and input process into the approach, contact, loosening, pushing-out, pushing-in, tightening and withdrawal stages according to the visual and force perception information and switching between the position control and force control. The task instruction module is used for receiving the "input" or "withdrawal" operation instruction of the pressing plate and providing the instruction to the segmented control module and selecting the pressing plate input process or the pressing plate withdrawal process according to different instructions. The visual verification module is used for collecting the image after the operation and judging whether the pressing plate withdrawal and input state is correct through the image comparison before and after the withdrawal and input.

2. The mobile manipulator press unloading operation system based on the segmented operation according to claim 1, characterized in that, The physical operation of the pressing plate withdrawal and input includes the pressing plate withdrawal operation and the pressing plate input operation.

3. The mobile manipulator press unloading operation system based on the segmented operation according to claim 1, characterized in that, The visual perception module adopts the depth camera to collect the color image and the depth image at the same time, and the four steps of image preprocessing, target identification, pixel coordinate extraction and three-dimensional coordinate back projection are as follows. The depth camera collects the RGB-D image set of the pressing plate area, the RGB image is used for the shape identification of the lock and the connecting piece, and the pixel coordinates of the pressing plate terminal post lock and the connecting piece are extracted through the visual algorithm; the depth image is used for obtaining the depth value; The pixel coordinates are back projected to the three-dimensional space of the camera coordinate system by using the pinhole model, and the three-dimensional points obtained according to the depth value represent the actual space position of the lock or the connecting piece in the camera coordinate system; The three-dimensional coordinates obtained by back projection and the identified target categories are output to the coordinate transformation module to provide data input for the subsequent coordinate conversion, trajectory planning and manipulator control, and the target categories include the lock or the connecting piece.

4. The mobile manipulator press unloading operation system based on segmented operation according to claim 1, characterized in that, The coordinate transformation module obtains the target pose by combining the lock nut or the direction feature of the continuous sheet recognized by the visual perception module with the camera pose rotation matrix performs transformation on the direction vector, and the target pose matrix is obtained from , wherein is a target orientation matrix in the camera coordinate system, and the lock nut or the continuous sheet direction feature is a lock nut normal vector direction or a continuous sheet edge direction.

5. The mobile manipulator press unloading operation system based on segmented operation according to claim 1, characterized in that, The specific process of the force perception module is as follows: Force sensors detect contact force, lateral force and torque signals in real time wherein is the three-axial contact force, is the three-axial torque; The original signal is offset compensated and proportionally calibrated by a sensor calibration matrix, so that the measured force meets a linear model , is a calibration gain matrix, is a zero bias vector; The force sensor coordinate system is converted to the tool coordinate system of the manipulator end or the base coordinate system of the manipulator; The force / torque signal is filtered in the time domain to eliminate high-frequency noise and obtain a smooth force signal; Based on the filtered force and torque signal, the segmented control module sets multiple judgment thresholds to distinguish the loosening, pushing-out, pushing-in and tightening stages.

6. The segmented operation based mobile manipulator press platen operation system of claim 1, wherein, The segmented switching control module controls the movement of the six-degree-of-freedom mobile manipulator in segments according to visual positioning, task instructions of the task instruction module and force feedback of the force perception module, and is divided into an approaching stage and a contact and operation stage, and the specific operation is as follows: In the approaching stage, the approaching path is planned according to the three-dimensional position and target pose provided by the coordinate transformation module, so that the end operation module of the six-degree-of-freedom mobile manipulator gradually approaches the lock nut or the connecting piece, and position control is adopted; In the contact and operation stage, when the end operation module enters the contact area, the force perception module starts to detect the contact force, lateral force and torque in real time, and after the contact force reaches the set threshold, the position control is switched to force control, the force control adopts an impedance control model, and the force perception is operated in real time in the control cycle, to provide a stage determination signal for the segmented switching control module, and when an abnormal force peak or torque is detected, a protection mechanism is triggered immediately, so that the six-degree-of-freedom mobile manipulator performs a retreat action.

7. The segmented operation based mobile manipulator press platen operation system of claim 6, wherein, The segmented switching control module sets multiple determination thresholds to distinguish the unscrewing, prying out, prying in and tightening stages, and the specific operation is as follows: In the unscrewing phase of the lock nut, a first contact force threshold is used The contact establishment is judged, and the unscrewing torque criterion is used at the same time Wherein is the unscrewing threshold, It is obtained by differentiating the continuous torque signal; In the continuous dialing phase, a second contact force threshold is used , displacement changes in the continuous dialing direction and the reaction force are monitored, and the dialing is completed to satisfy a joint criterion , where is a displacement threshold, is a reaction force threshold. In the connecting piece prying in stage, whether the connecting piece has entered the closed position is judged by monitoring the change trend of the prying in force with displacement, and when the force change tends to be stable and the displacement reaches the set position, it is determined that the prying in is completed; In the lock nut tightening phase, a third contact force threshold is used Contact is established, a torque convergence criterion is used to determine that the lock nut tightening is complete, wherein is a tightening threshold.

8. The segmented operation based mobile manipulator press platen operation system of claim 1, wherein, The task instruction module analyzes the input "switch in / out" command into a Cmd signal, that is, Cmd=Strap_Enable or Cmd=Strap_Disable, and then transmits the Cmd to the segmented switching control module, so that the segmented switching control module determines whether to perform the pressing plate exit operation or the pressing plate entry operation.

9. A method for switching on and off a mobile robot arm presser based on the segmented operation according to any one of claims 1-8, characterized in that, The method comprises the following steps: Step S1: collecting the pressing plate area image and identifying the pixel positions of the pressing plate terminal post lock nut and connecting piece; Step S2: converting the pixel positions into manipulator base coordinates; Step S3: the task instruction module analyzes the instruction to generate the pressing plate entry command Cmd=Strap_Enable or the pressing plate exit instruction Cmd=Strap_Disable; Step S4: according to the generated instruction Cmd, selecting the exit operation or the entry operation and planning the approaching trajectory of the six-degree-of-freedom mobile manipulator; Step S5: detecting the contact state through the force perception module, and switching to force control when the contact force reaches the set threshold; Step S6: when Cmd=Strap_Disable, the segmented switching control module is completed for the lock nut unscrewing and the connecting piece prying out; Step S7: when Cmd=Strap_Enable, the segmented switching control module is completed for the connecting piece prying in and the lock nut tightening; Step S8: after the operation is completed, the pressing plate switching state is judged through the visual verification module; Step S9: when the verification fails, the trajectory of the six-degree-of-freedom mobile manipulator is re-planned, and steps S3 to S7 are repeated.

Images