Cooperative control system and robot for high-precision heavy-duty disassembly

Abstract

The invention provides a cooperative control system and robot for high-precision heavy disassembly, and relates to the technical field of engineering machinery automation. The cooperative control system comprises a main controller, a parallel mechanical arm linear walking module, a contact type vertical tool setting module and a laser ranging module. The parallel mechanical arm linear walking module drives the cutting execution tail end to conduct high-rigidity linear feeding along a preset path through a space parallel closed-loop connecting rod mechanism, the contact type vertical tool setting module obtains vertical reference data through physical contact, and the laser ranging module monitors distance data in real time. The main controller is configured to calibrate the initial attitude based on the vertical reference data and correct the feed depth in real time according to the distance data. The robot adopts a crawler-type walking chassis and a sunken electrical cabinet, so that the problems of insufficient rigidity and poor environmental adaptability of heavy disassembly are solved.

Description

Technical Field

[0001] This invention relates to the field of automation technology for engineering machinery, and in particular to a collaborative control system and robot for high-precision heavy dismantling. Background Technology

[0002] With the upgrading of new energy industry infrastructure, the demand for on-site harmless treatment of large composite material components such as waste wind turbine blades is becoming increasingly urgent. Such dismantling operations are typically carried out in complex terrain and variable climate environments, and the objects being handled are characterized by large geometric dimensions, high material strength, and significant internal residual stress. Currently, most mobile dismantling equipment uses a traditional engineering machinery chassis equipped with a tandem articulated robotic arm, and is operated from a separate external control room.

[0003] In actual heavy dismantling processes, traditional tandem robotic arm structures have relatively limited rigidity when extended over long distances. When the cutting end is subjected to enormous cutting reaction forces, the long cantilever structure is prone to low-frequency chatter, which affects the straightness of the cutting trajectory and the smoothness of the cut to some extent. Meanwhile, the surfaces of objects to be dismantled in the field are often covered with weathered materials or soil, and their geometry is irregular. The high-density dust accompanying the cutting process can obstruct the view of optical sensors, making it difficult to establish accurate vertical tool alignment and depth tracking benchmarks solely based on vision or manual remote control. Furthermore, existing equipment mostly relies on deploying hydraulic support legs to maintain vehicle balance. This deployment process is relatively cumbersome and limits the equipment's maneuverability and accessibility in narrow gaps in rubble. Summary of the Invention

[0004] The purpose of this invention is to provide a collaborative control system and robot for high-precision heavy disassembly, in order to solve the problems pointed out in the background art.

[0005] In a first aspect, the present invention provides a collaborative control system for high-precision heavy disassembly, the collaborative control system comprising: a main controller, a parallel robotic arm linear motion module, a contact-type vertical tool setting module, and a laser ranging module; The parallel robotic arm linear travel module is used to drive the cutting execution end to perform high-rigidity linear feed along a preset dismantling path; The contact-type vertical tool setting module is located at the cutting execution end and is used to obtain vertical reference data of the surface of the object to be disassembled by physical contact before the disassembly operation. The laser ranging module is located at the cutting execution end and is used to monitor the distance data between the cutting execution end and the surface of the object to be disassembled in real time. The main controller is configured to calibrate the initial attitude based on the vertical reference data and correct the feed depth of the parallel robotic arm linear travel module in real time according to the distance data, so as to coordinate the control of the disassembly operation.

[0006] Optionally, the parallel robotic arm linear travel module adopts a spatial parallel closed-loop linkage mechanism. The main controller controls the extension and retraction of each branch of the spatial parallel closed-loop linkage mechanism through inverse kinematics calculation, and uses the structural stiffness of the parallel mechanism to counteract the reaction force during the heavy disassembly process.

[0007] Optionally, the contact-type vertical tool setting module is equipped with a pressure-sensing probe; The vertical calibration logic of the main controller is as follows: control the parallel robotic arm linear walking module to drive the pressure sensing probe to slowly feed towards the surface of the object to be disassembled, and when a contact signal is received, lock the current pose coordinates as the vertical tool setting zero point.

[0008] Optionally, the main controller is equipped with a dynamic depth compensation module, which is used to compare the real-time distance data fed back by the laser ranging module with the preset cutting depth. When surface undulations are detected, the compensation amount of the parallel robotic arm linear travel module in the normal direction is adjusted in real time.

[0009] Optionally, it also includes an environmental sensing and dust removal control unit, which is used to control the dust removal hood covering the cutting area to activate the negative pressure dust suction function according to the working status of the parallel robotic arm linear walking module.

[0010] Optionally, the data interaction between the parallel robotic arm linear walking module, the contact vertical tool setting module, and the laser ranging module is all converged to the main controller via an industrial bus. The main controller is configured with an adaptive motion algorithm for unstructured outdoor environments.

[0011] Optionally, the main controller is also configured with multi-point fitting logic, which controls the contact-type vertical tool setting module to collect the coordinates of at least three non-collinear points on the surface of the object to be disassembled, fits the normal vector of the cutting plane, and corrects the entry angle of the cutting tool.

[0012] Optionally, the main controller also integrates a sensor confidence dynamic reconfiguration module based on load gradient, which is configured to execute the following control logic: The rate of change of the driving current of the parallel robotic arm linear travel module is monitored in real time, and the rate of change of the driving current is used as a feedforward parameter to determine the intensity of dust outbreak in the cutting area. When the rate of change of the driving current exceeds a preset threshold, the main controller determines that it is a high dust and strong interference condition. At the same time, it triggers the environmental perception and dust removal control unit to start the negative pressure dust collection function and automatically reduces the confidence weight of the distance data fed back by the laser ranging module in the closed-loop control. Meanwhile, the main controller increases the confidence weight of the inverse kinematics-derived depth data based on the parallel robotic arm linear walking module, and uses the inverse kinematics-derived depth data to temporarily replace the distance data to maintain the cutting path until the rate of change of the driving current falls back to the steady-state range, so as to avoid laser ranging distortion caused by dust scattering during heavy disassembly.

[0013] Optionally, the main controller is further configured with a macro-micro coordinated stress deadlock relief module, which is configured as follows: During the disassembly process, the load status of the cutting execution end is monitored in real time. When a sudden increase in load is detected and the feed rate is forced to drop to zero, it is determined that a clamping deadlock caused by material stress release has occurred. The main controller immediately sends a macro-axis locking command to the parallel robotic arm linear walking module to force it to maintain its current position in the linear walking direction without moving backward, so as to serve as a stable mechanical support base. Simultaneously, the main controller controls the spatial parallel closed-loop linkage mechanism to enter the high-frequency micro-motion mode, driving the cutting execution end to perform micro-amplitude high-frequency reciprocating vibration in the tangential plane perpendicular to the feed direction. The high-frequency response characteristics of the parallel mechanism are used to loosen the deadlock contact surface. After the load monitoring value decreases, the cutting execution end is controlled to perform a micro-retraction action.

[0014] In a second aspect, the present invention provides a robot for high-precision heavy dismantling, comprising a tracked chassis, a sunken electrical cabinet, an integrated host computer, and a collaborative control system as described in any one of the first aspects; The parallel robotic arm linear travel module is mounted on the tracked chassis and is used to perform operations without hydraulic support legs. The recessed electrical cabinet is embedded in the internal space of the tracked chassis, and the integrated host computer is directly mounted on the robot body for human-machine interaction and issuing control commands.

[0015] The present invention has achieved the following beneficial effects: This invention utilizes the high structural rigidity of a spatial parallel closed-loop linkage mechanism to construct a parallel robotic arm linear travel module, effectively decomposing the huge reaction force during heavy disassembly into axial loads of each branch, improving the stress state of traditional series cantilever structures under heavy cutting conditions, thereby ensuring the straightness of the cutting trajectory and the quality of the cut.

[0016] This invention solves the problem of establishing benchmarks on complex surfaces in the field by combining contact detection and non-contact ranging. In particular, the dynamic reconfiguration logic of sensor confidence based on load gradient can automatically reduce the weight of optical data and switch to inverse kinematics deduction mode the instant that dust bursts interfere with laser ranging. Combined with the linkage of environmental perception and dust removal unit, it ensures the continuity of operation and control stability in harsh visibility environments.

[0017] The macro-micro coordinated stress deadlock release mechanism designed in this invention utilizes the high-frequency micro-motion response capability of a parallel mechanism to address the tool clamping phenomenon commonly encountered in the disassembly of large components. When deadlock occurs, the system can lock the macro axis as a stable support and disrupt the static friction state through tangential high-frequency vibration of the micro axis, achieving autonomous escape while protecting the tool and drive motor.

[0018] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of a collaborative control system for high-precision heavy disassembly in an embodiment of the present invention. Detailed Implementation

[0021] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0022] This invention provides a collaborative control system and robot for high-precision heavy dismantling. This technical solution is specifically designed for on-site dismantling of large composite material components (such as discarded wind turbine blades) in unstructured outdoor environments. In such conditions, the workpiece typically has irregular geometry, severe surface weathering and corrosion, high internal residual stress, and dust pollution during the cutting process. Traditional engineering machinery or general industrial robots based on serial robotic arms are insufficient in terms of rigidity, precision, and environmental adaptability.

[0023] In this embodiment, the robot used for high-precision heavy dismantling employs an integrated tracked chassis as its mobile base. To adapt to the common terrain of gravel, mud, and slopes at dismantling sites, the tracked chassis uses wide, high-grip engineering rubber and steel wire composite tracks, and is equipped with an independent hydraulic drive unit or a high-torque permanent magnet synchronous motor drive unit, ensuring that the robot has the ability to climb slopes and turn on the spot under full load.

[0024] The recessed electrical cabinet is not simply mounted on the vehicle body surface, but is embedded in the internal cavity of the tracked chassis, specifically between the two track beams. The main controller of this invention's collaborative control system is integrated within the recessed electrical cabinet. The main controller preferably employs a high-performance motion controller based on a heterogeneous architecture of an industrial-grade PC and a field-programmable gate array (FPGA), or an embedded controller based on a real-time operating system. The main controller connects to an integrated host computer located on the side of the robot body via gigabit Ethernet or fiber optic cable. The integrated host computer is equipped with a high-brightness industrial touchscreen and ruggedized physical buttons for operators to perform human-machine interaction, set operating parameters, monitor 3D status, and perform emergency stop operations on-site. Furthermore, the integrated host computer also has a built-in wireless communication module, supporting remote mapping operations by operators in a safe area via a handheld terminal.

[0025] The parallel robotic arm linear travel module is mounted on the upper surface platform of the tracked chassis. To address the issues of insufficient stiffness and low-frequency vibration at the end of traditional serial robotic arms (such as six-axis articulated arms) in long cantilever states, which can lead to skeletal deviations, the parallel robotic arm linear travel module in this embodiment employs a high-stiffness spatial parallel closed-loop linkage mechanism. The specific configuration of this mechanism can preferably be 3-RPS (revolute-prismatic-spherical joint), 6-UPS (universal joint-prismatic-spherical joint), or a hybrid mechanism with redundant degrees of freedom. Regardless of the specific topology used, its core feature is that the cutting execution end (i.e., the moving platform) is connected to the base (fixed platform) via at least three independent telescopic branches, forming a stable spatial truss structure. In heavy dismantling operations, when the cutting tool cuts into the material and is subjected to enormous normal repulsive force and tangential resistance, these forces do not accumulate bending moments as in serial arms, but are decomposed into axial tensile and compressive forces of each branch. Because the axial stiffness of the linkage is much greater than its bending stiffness, the cutting execution end exhibits extremely high structural stability. The parallel robotic arm linear motion module also includes a heavy-duty linear guide module, which drives the entire parallel mechanism to perform a wide range of linear feeds along the longitudinal direction of the vehicle body.

[0026] The electronic and electrical architecture of the collaborative control system is built on a high-speed industrial fieldbus, preferably using the EtherCAT or CANopen bus protocol. The main controller acts as the bus master, and is connected in a daisy-chain topology to the servo driver of the parallel robotic arm linear motion module, the I / O acquisition node of the contact vertical tool setting module, the communication gateway of the laser ranging module, and the frequency converter of the environmental perception and dust removal control unit.

[0027] Furthermore, the main controller internally operates an inverse kinematics calculation engine for spatial parallel closed-loop linkage mechanisms. During operation, the main controller plans a series of discrete pose points at the cutting execution end in a Cartesian coordinate system based on a preset disassembly path (e.g., a spatial straight line). For each pose point, the inverse kinematics engine calculates the target displacement of each branch drive unit (e.g., an electric cylinder) in real time based on the geometric parameter model of the mechanism. Notably, to cope with the nonlinear friction and gravity effects caused by heavy loads, the main controller, while sending position commands, also calculates feedforward torque commands based on a rigid body dynamics model, directly superimposing them into the current loop setpoint of the servo driver. This dynamic feedforward control strategy enables the parallel mechanism to maintain extremely high trajectory tracking accuracy even when subjected to cutting reaction forces of hundreds of kilograms. It effectively utilizes the structural stiffness of the parallel mechanism to offset the reaction forces during heavy disassembly, suppressing low-frequency chatter caused by cutting impact, thereby ensuring the straightness of the cutting trajectory and the smoothness of the cut.

[0028] However, high-precision motion control requires an accurate workpiece coordinate system. In outdoor environments, materials to be dismantled (such as discarded wind turbine blades) are often randomly piled up, their surfaces covered with oil, rust, snow, or mud, and may themselves be warped or deformed. Traditional non-contact vision sensors are prone to failure in such environments, making it impossible to establish a reliable working reference. Therefore, this embodiment integrates a contact-type vertical tool setting module at the cutting execution end. This module is equipped with a highly sensitive pressure-sensing probe, which can be a single-axis force sensor based on the principle of resistance strain gauges or a telescopic probe integrated with a microswitch. The probe's installation position is precisely calibrated, and its trigger point has a defined geometric offset relationship with the cutting tool's center of action in the mechanical coordinate system.

[0029] Before the disassembly operation officially begins, the system executes an automated initialization process. Specifically, firstly, the main controller controls the parallel robotic arm's linear motion module to move the cutting end effector to a preparatory height of approximately 50mm to 100mm above the target cutting area. Then, the robotic arm switches to probing mode, driving the pressure-sensing probe at a preset slow speed (e.g., 2mm / s to 5mm / s) towards the surface of the object to be disassembled. During this process, the main controller activates a high-speed interrupt service to monitor the probe's feedback signal in real time. To prevent false triggering due to wind or equipment vibration, the main controller incorporates signal debouncing logic; valid contact is only determined when the contact signal remains valid for a preset time threshold (e.g., 20ms). Once valid contact is determined, the main controller uses the bus's distributed clock function to instantly latch the current encoder position values ​​of each axis and, combined with forward kinematics calculations, calculates the precise three-dimensional coordinates of the current contact point in the robot's base coordinate system. These coordinates are then locked as the vertical tool zero point, serving as the absolute reference for all subsequent depth control. To further protect the probe, upon triggering, the main controller will immediately control the robotic arm to perform a flexible retraction motion, raising it to a safe height.

[0030] Considering that the surface of the object to be disassembled may not be horizontal but has a certain tilt angle, a single tool setting point is insufficient to describe the spatial attitude of the surface. Therefore, the main controller is also equipped with multi-point fitting logic. During the initialization phase, the main controller controls the parallel robotic arm to drive the contact-type vertical tool setting module to sequentially collect the contact coordinates of at least three non-collinear points (e.g., three points distributed in an equilateral triangle) within a small neighborhood around the cutting starting point. Based on the coordinate data of these three points, the mathematical processing unit inside the main controller uses spatial analytical geometry algorithms or the least squares method to fit the equation of the cutting plane and then calculates the normal vector of the plane. Based on this normal vector, the main controller calculates the spatial attitude angle that the cutting tool should have, and adjusts the attitude of the moving platform by controlling the differential extension and retraction of each branch of the parallel robotic arm, so that the axis of the cutting tool and the normal vector of the workpiece surface maintain a preset process angle (e.g., vertical or a specific bevel angle), thereby correcting the entry angle of the cutting tool. This process not only avoids the risk of tool jamming caused by tool tilting but also significantly extends the tool's service life.

[0031] After establishing the operational baseline and initial posture, the disassembly operation enters the dynamic cutting phase. Due to the high-speed rotation and severe vibration of the cutting tool during cutting, the contact probe must be retracted or removed, and the task of maintaining depth is handled by the laser ranging module. The laser ranging module is positioned at the cutting end, ahead of the tool's feed direction, serving as a forward-looking detection mechanism. This module emits a laser beam and receives diffuse reflection echoes from the surface, monitoring the relative distance between the cutting end and the surface of the object to be disassembled in real time at a high frequency (e.g., 1 kHz).

[0032] The main controller has a built-in dynamic depth compensation module, which works as follows: During the cutting process, the laser ranging module continuously transmits real-time distance data. The main controller compares this data with the preset ideal cutting height (usually determined by process requirements, such as maintaining a nozzle distance of 10mm from the surface) to calculate the height deviation. It's important to note that because there is a physical distance between the laser ranging point and the actual cutting point, the undulations measured by the laser are actually the terrain the cutting point will encounter at future moments. Therefore, the dynamic depth compensation module maintains a time-based first-in-first-out queue. The system calculates the delay time based on the current feed rate and physical distance, and pushes the height deviation data into the queue. When the cutting tool actually reaches the measured position, the main controller retrieves the corresponding height deviation value from the queue and adds it to the Z-axis motion command, driving the parallel robotic arm's linear motion module to perform a slight compensation motion in the normal direction. This spatiotemporal aligned look-ahead compensation strategy allows the robot to accurately follow the undulations of the workpiece surface for equal-depth cutting, avoiding problems such as incomplete cuts or excessive cutting depth damaging the base due to surface undulations.

[0033] Heavy dismantling, especially when cutting fiberglass, concrete, or corroded metal, generates large amounts of high-density dust. This dust not only pollutes the environment but, more seriously, obstructs the optical path of the laser ranging module, causing distance data distortion (e.g., the laser hitting a dust cloud might be mistaken for a closer distance). The traditional approach is to keep the dust collection equipment running continuously, but this is energy-intensive and often reacts slowly to bursts of dust. This system introduces an environmental sensing and dust collection control unit and implements dynamic reconfiguration logic for sensor confidence based on load gradients.

[0034] Preferably, the environmental sensing and dust removal control unit includes a follow-up dust hood covering the cutting area, a high-power negative pressure fan, and connecting pipelines. The follow-up dust hood maintains contact with the workpiece surface through a flexible skirt, forming a relatively enclosed cutting chamber. The main controller senses the operating status by monitoring the drive current of the parallel robotic arm's linear travel module. Specifically, the main controller integrates a sensor confidence dynamic reconfiguration module based on load gradient. This module calculates the rate of change of the drive current of the cutting spindle motor or feed axis motor in real time. When the cutting tool cuts into the reinforcing ribs, sand layers, or encounters stress release within the material, the cutting resistance increases instantaneously, causing drastic fluctuations in the drive current; and this drastic cutting process is often accompanied by explosive dust generation. Therefore, the rate of change of the drive current is used as a feedforward parameter to determine the intensity of dust explosions in the cutting area.

[0035] When the main controller detects that the rate of change of the drive current exceeds a preset threshold, it determines that the system has entered a high-dust, high-interference operating condition. At this point, the main controller immediately executes two actions: First, it sends a full-speed operation command to the environmental perception and dust removal control unit via the bus, instantly accelerating the negative pressure fan to its maximum speed and forcefully sucking the dust into the dust collection hood before it can spread. Second, at the control algorithm level, the main controller automatically reduces the confidence weight of the distance data fed back by the laser ranging module in the closed-loop control. For example, in the Kalman filter algorithm, the covariance matrix value of the laser measurement noise is increased, reducing its weight from 0.9 to 0.1 or even zero. Simultaneously, the main controller increases the confidence weight of the inverse kinematics-derived depth data based on the parallel robotic arm's linear motion module. Inverse kinematics-derived depth data refers to the theoretical position calculated by the controller using a mechanical system model based on the reliable position from the previous moment and the current motion command. In the short term, the inertia of the mechanical system ensures the reliability of the theoretical position. The system uses inverse kinematics-derived depth data to temporarily substitute for failed laser data to maintain the cutting path, i.e., it enters a blind cutting sustaining mode until the current change rate falls back to the steady-state range, indicating smooth cutting, dust dissipation, and reliable recovery of laser readings. Then, the weights of the laser data are smoothly restored. This dynamic reassignment strategy effectively solves the reliability problem of optical sensors under harsh working conditions, ensuring the continuity of operations.

[0036] Furthermore, to address the common "stress lock-up" or "tool clamping" phenomena in heavy dismantling, the main controller is equipped with a macro-micro coordinated stress lock-up release module. When dismantling large, stressed components, as the kerf extends, the release of residual stress in the material may cause the kerf to close, tightly clamping the tool. Forcibly retracting the tool at this point can easily lead to tool breakage or motor overload. The stress lock-up release module is configured to execute release logic: during the dismantling operation, it monitors the load status (such as current or torque) at the cutting end in real time. When a sudden increase in load is detected and the feed rate is forced to drop to zero (position error exceeds limits), a lock-up is determined to have occurred. The main controller immediately sends a macro-axis locking command to the parallel robotic arm linear motion module, forcing the servo drive into a high-rigidity position-holding mode, locking the current feed position, and preventing the mechanism from springing open under reaction force.

[0037] Next, the main controller controls the spatial parallel closed-loop linkage mechanism to enter high-frequency micro-motion mode. Utilizing the high bandwidth of the parallel mechanism, it drives the cutting end effector to perform micro-amplitude (e.g., 0.5mm amplitude), high-frequency (e.g., 30Hz) reciprocating vibration in a tangential plane perpendicular to the feed direction (i.e., the side of the tool). This micro-vibration effectively breaks down the static friction between the kerf and the tool, and pulverizes locally stuck debris. While maintaining vibration, the main controller monitors load changes; once a load decrease is detected, it attempts to superimpose a slight retraction command until the tool is safely freed.

[0038] This invention provides a robot for high-precision heavy disassembly, which includes a tracked chassis, a sunken electrical cabinet, an integrated host computer, and the aforementioned collaborative control system.

[0039] In heavy-duty dismantling operations, especially during the cutting of large wind turbine blades (typically made of fiberglass reinforced composites), the cutting reaction force often exhibits characteristics of high load (up to thousands of Newtons), low-frequency impact (frequency typically between 1-10Hz), and abrupt changes in direction. Traditional wheeled or light-duty tracked chassis, lacking hydraulic outrigger support, are prone to slippage or elastic vibrations due to tire elastic deformation or chassis suspension redundancy. These minute base movements are amplified by the overhead parallel robotic arm, directly causing the end-cutting trajectory to deviate, and even resulting in blade breakage.

[0040] Therefore, this embodiment of the robot uses a tracked chassis frame with an all-steel welded box-type structure. The main beam material is preferably high-strength low-alloy structural steel (such as Q345E or Q460), and undergoes overall annealing heat treatment to eliminate residual welding stress, ensuring impact toughness and dimensional stability in a wide-temperature environment ranging from -40℃ to +50℃. The walking drive system uses dual independent low-speed, high-torque hydraulic motors, or permanent magnet synchronous motors (PMSM) with multi-stage planetary reducers. The tracks are made of engineering rubber with a high-grip herringbone pattern, internally vulcanized with high-strength steel cords to prevent tensile deformation; or, for extremely harsh terrain, steel track plates with single or double teeth are used.

[0041] To enable operation without hydraulic outriggers, the chassis integrates a high-preload automatic tensioning and locking mechanism. When the robot reaches the work position and stops moving, the main controller sends a command to lock the electromagnetic brake of the walking motor and raises the track tension from the low threshold of the walking mode (e.g., 5 MPa) to the high threshold of the working mode (e.g., 12 MPa) via the hydraulic tensioning cylinder. This action eliminates mechanical backlash in the walking wheel system (drive wheels, guide wheels, support wheels). As a result, a large rigid contact area is formed between the tracked chassis and the ground. Utilizing the static friction generated by the robot's own weight (typically designed between 3 and 10 tons) and the ground, a virtual gravity anchor point is constructed sufficient to resist cutting reaction forces. This design not only eliminates the need for bulky hydraulic outriggers and reduces deployment and retraction time, but also gives the robot extremely high mobility in narrow gaps in rubble.

[0042] Secondly, this embodiment utilizes the underbody space between the track beams on both sides of the tracked chassis to design an embedded electrical compartment. The outer shell of this sunken electrical cabinet is welded from double-layer armored steel plates, providing bulletproof impact resistance, and is filled with heat-insulating and flame-retardant rock wool or aerogel felt in the middle.

[0043] In terms of thermal management, considering the potential heat dissipation issues arising from embedded installation, the electrical cabinet is designed with an independent forced air cooling or phase change liquid cooling circulation system. The air intake is located high on the rear side of the vehicle body and is equipped with a cyclone pre-filter and a HEPA high-efficiency filter to prevent conductive carbon fiber dust or metal dust from the disassembly site from entering the cabinet and causing short circuits. Furthermore, to isolate the precision electronic components (such as the main controller FPGA chip and the signal processing board of the laser ranging module) from road impacts during travel and high-frequency cutting vibrations during operation, the sunken electrical cabinet is not rigidly bolted to the chassis frame, but is suspended by a set (usually 4 to 8) of wire rope vibration isolators. These wire rope vibration isolators utilize the dry friction damping characteristics between the strands of the wire rope to simultaneously attenuate vibration energy in three translational directions (X / Y / Z) and three rotational directions (Roll / Pitch / Yaw), ensuring that the vibration acceleration of sensitive components such as the main controller inside the electrical cabinet is always controlled within a safe range (e.g., less than 0.5g).

[0044] Finally, the integrated host computer, serving as the human-machine interface, is directly located in a protected recessed area on the side or rear of the robot body. This host computer is a military-grade rugged computer meeting IP67 protection standards. Its display uses a high-brightness screen with optical bonding technology, ensuring clear readability even under direct sunlight in the field. The integrated host computer communicates with the main controller in the recessed electrical cabinet via an internal high-speed bus. It is responsible not only for displaying the cutting trajectory, sensor status, and alarm information, but also for handling some non-real-time data processing tasks, such as storing historical operation data, generating dismantling logs, and uploading status data to the remote dispatch center via 4G / 5G / BeiDou short message networks. To prevent accidental operation, a hard-wired emergency stop button and a three-position mode selection switch (manual / teach / automatic) are also provided next to the host computer, forming a double safety redundancy.

[0045] In the field of modern high-end equipment control, the real-time and synchronous nature of data transmission is crucial to determining control accuracy. The collaborative control system of this invention employs Ethernet-based industrial fieldbus technology, preferably the EtherCAT (Ethernet for Control Automation Technology) real-time industrial Ethernet bus.

[0046] The data exchange between the parallel robotic arm linear motion module (containing multiple servo drives), the contact-type vertical tool setting module (containing high-speed digital I / O), the laser ranging module (containing a gateway interface), and the environmental perception and dust removal control unit (containing a frequency converter) is all converged to the main controller via this industrial bus. The main controller, acting as the bus master, ensures that the clock synchronization error of all slave nodes in the network is less than 1 microsecond through a "processing on the fly" transmission mode.

[0047] For example, during the stress lockout release process in the aforementioned embodiment, the main controller needs to simultaneously freeze the motion of the macroscopic linear axis and initiate high-frequency vibration of the microscopic parallel mechanism the instant (milliseconds) upon detecting a sudden load change. If there is a significant delay or jitter in the bus communication, the macroscopic axis may continue to advance under inertia, leading to tool breakage or workpiece damage. Using EtherCAT's Distributed Clocks technology, the main controller can pre-set the absolute moment of action execution, ensuring that all axes respond to commands within the same microsecond.

[0048] The adaptive motion algorithm configured inside the main controller for unstructured outdoor environments is the core software of this system. This algorithm framework comprises a perception layer, a decision-making layer, and an execution layer.

[0049] At the perception layer, the algorithm not only receives distance data from the laser ranging module and switching signals from the contact probe, but also integrates current loop data (reflecting torque), speed loop data, and data from the servo motors of each branch of the parallel robotic arm, as well as data from the chassis tilt sensor (IMU). The algorithm introduces an extended Kalman filter (EKF) state observer to fuse this multi-source heterogeneous data. For example, when the robot operates on soft ground, the chassis may experience slight settlement or tilting. The adaptive algorithm senses the low-frequency drift of the chassis attitude through the IMU and, combined with the changing trends of the laser ranging data, distinguishes between the terrain undulations and chassis settlement components in real time, thus enabling targeted compensation in subsequent control.

[0050] At the decision-making level, the adaptive motion algorithm dynamically adjusts its operational strategy based on the local environment model constructed by the perception layer (including workpiece surface geometry, material hardness distribution estimation, and current dust concentration). In addition to the aforementioned blind-cutting sustaining mode and deadlock release mode, the algorithm also includes a variable impedance control strategy. At the instant the cutting tool contacts the workpiece surface (tool entry phase), the algorithm controls the parallel mechanism to exhibit low stiffness (low impedance) to compliantly adapt to the contact impact; once the cut is stable, it immediately switches to a high-stiffness (high impedance) position control mode to ensure trajectory accuracy; when cutting to the edge of the workpiece and about to cut off (tool exit phase), the algorithm readjusts the impedance parameters to prevent damage to the robotic arm from a "missed" impact caused by material fracture.

[0051] At the execution layer, the algorithm maps the planned Cartesian space trajectory to the joint space through inverse kinematics calculation. Considering the kinematic coupling and nonlinear characteristics of parallel mechanisms, the algorithm incorporates a built-in dynamic model compensator. This compensator, based on the Newton-Euler equations or the Lagrange equations, calculates in real time the changes in gravity components, Coriolis force, and centrifugal force caused by the attitude changes of the moving platform, and converts them into feedforward torque commands to be sent to the servo drive.

[0052] In actual heavy-duty dismantling operations, there is a complex nonlinear relationship between dust generation and laser sensor failure. Traditional control systems often use simple threshold judgments (e.g., stopping operation when the laser reading changes abruptly), which causes the robot to frequently stop in dusty environments, making it impossible to operate. This invention proposes a dynamic reconfiguration strategy based on physical feedforward.

[0053] Specifically, the sensor confidence dynamic reconfiguration module in the main controller includes three sub-units: load gradient monitoring unit, confidence calculation unit, and trajectory fusion unit.

[0054] The load gradient monitoring unit collects the drive current of the parallel robotic arm linear motion module in real time. Since the motor's output torque is directly proportional to the current, and torque directly reflects cutting resistance, a sharp increase in cutting resistance usually indicates that the tool has cut into a material region that generates a large amount of dust (such as a concrete interlayer or a layer of corroded metal). This unit calculates the first derivative (rate of change) of the current. It is then low-pass filtered to remove electromagnetic noise.

[0055] The confidence calculation unit receives the filtered rate of change of current and uses a preset membership function (such as the sigmoid function or a piecewise linear function) to calculate the confidence weight of the laser data. .when When the cutting speed is low (smooth cutting). Close to 1; when When the preset threshold is exceeded (indicating an impending dust explosion), It decays rapidly to 0. It is worth noting that, in order to prevent control oscillations, the weight recovery process (from 0 to 1) is designed to be a slow process with hysteresis characteristics. That is, the weight of the laser data is gradually restored only after the current change rate has been low for a long time and the laser echo intensity (RSSI) has returned to normal.

[0056] The trajectory fusion unit is responsible for generating the final Z-axis height control command. The calculation formula is as follows: ; in, These are the measured values ​​from the laser ranging module. The theoretical position value is calculated based on the previous position, the current speed command, and the inverse kinematics model of the robotic arm.

[0057] Through this dynamic reconfiguration mechanism, when a dust outburst causes spurious signals in the laser ranging value (such as a decrease due to scattering or an increase due to obstruction), because at this time... The laser data has been suppressed by the current feedforward signal, so erroneous laser data will not affect the final control command. The system smoothly transitions to model predictive control mode, using the inertia of the mechanical system to maintain the cutting path. After the dust is sucked away by the dust removal unit and the operating conditions stabilize, the system seamlessly switches back to laser closed-loop control mode.

[0058] When disassembling large components with high residual stress, "tool jamming" is a major cause of equipment damage. In traditional pure position control mode, when tool jamming occurs, the position error increases, the integral term accumulates, leading to saturation of the motor output torque, and eventually triggering an overcurrent alarm and shutdown. At this point, the tool is often already stuck and difficult to remove.

[0059] The stress lock-out release module of the present invention utilizes the complementary characteristics of a parallel robotic arm linear travel module (macro-axis, large stroke, high stiffness but low frequency response) and a spatial parallel closed-loop linkage mechanism (micro-axis, small stroke, high frequency response).

[0060] When the main controller detects signs of tool lock-up (speed is zero and torque exceeds the limit), it immediately initiates the release process: Step 1: Macroscopic Axis Rigid Locking. The main controller sends a full closed-loop position lock command to the linear motion module's driver and temporarily increases the position loop gain (Kp), making the linear axis a highly rigid base. This is to prevent the reaction force from causing the entire robotic arm to retreat during subsequent vibrations, which could lead to tool breakage.

[0061] Step 2: Microscopic shaft impedance modulation and high-frequency excitation. The main controller switches the control mode of the parallel mechanism from position mode to impedance control mode, setting the stiffness in the tangential direction (i.e., the direction in which the tool is clamped) to a lower value, while maintaining high stiffness in the normal direction (feed direction). At the same time, the main controller superimposes a high-frequency sinusoidal torque command (e.g., 30Hz-50Hz, with an amplitude of 20%-30% of the rated torque) in the tangential direction.

[0062] Step 3: Resonance Loosening and Tool Retraction. High-frequency torque excitation causes the tool to generate micro-vibrations at the micrometer level within the kerf. This vibration effectively reduces the static friction coefficient between the tool and the material (physically, the dynamic friction coefficient is much smaller than the static friction coefficient) and pulverizes tiny debris at the jamming point. While maintaining vibration, the main controller attempts to superimpose a small tool retraction speed command. Once a decrease in load torque is detected (indicating loosening), the tool retraction speed is gradually increased until the tool is completely freed.

[0063] This embodiment, in conjunction with the above modules, provides the complete workflow and operation steps of the present invention in actual operation, specifically including: Step 1: Rapid Deployment and Self-Check. The operator drives the robot to the vicinity of the large component to be disassembled (such as the root of a wind turbine blade). After the robot stops, the main controller automatically executes the chassis locking program, driving the hydraulic cylinders to tension the tracks and lock the travel brakes. At the same time, the system performs a power-on self-check to confirm that the status of each servo axis is normal, the laser ranging module communication is normal, and the contact probe reset signal is valid.

[0064] Step Two: Macroscopic Positioning and Benchmark Establishment. The operator manually controls the parallel robotic arm's linear movement module via an integrated host computer or remote control terminal to move the cutting end effector above the cutting start point. The automatic tool setting program is initiated, and the main controller lowers the robotic arm, using pressure-sensing probes from the contact-type vertical tool setting module to perform multi-point contact detection around the cutting start point. Based on the three-dimensional coordinates of the contact points, the system fits the local surface equation and calculates the normal vector and zero vertical height of that point. The main controller, based on the calculated normal vector, drives the parallel mechanism to adjust the moving platform's posture, ensuring the cutting tool axis is perpendicular to the workpiece surface. This step ensures the accuracy of the initial posture in unknown terrain.

[0065] Step 3: Collaborative Cutting Operation. The system enters automatic operation mode. The main controller plans a linear feed trajectory based on the preset cutting length and speed. The parallel robotic arm linear travel module begins to drive the entire mechanism forward along the guide rail (macroscopic motion). Simultaneously, the laser ranging module scans the road surface height changes in real time and sends the data to the dynamic depth compensation module. The main controller, combined with the micro-motion capability of the parallel mechanism, adjusts the height of the cutting end in real time (Z-axis micro-adjustment) to ensure a constant cutting depth. During this process, macroscopic linear motion and microscopic normal compensation are performed simultaneously, achieving efficient and precise cutting.

[0066] Step Four: Disturbance Resistance and Adaptive Adjustment. During the cutting process, if a strong dust storm occurs, causing drastic fluctuations in laser readings and exceeding the driving current change rate limit, the main controller automatically triggers the sensor confidence dynamic reconfiguration logic, temporarily shielding the laser data, maintaining the trajectory through inverse kinematics deduction, and activating the dust removal unit at full speed. If material jamming occurs, obstructing the feed, the main controller automatically triggers the stress lockout release module, pausing macroscopic feed, initiating microscopic high-frequency vibration to loosen the tool, and attempting a slight tool retraction. These adaptive adjustments are entirely autonomously completed by the control algorithm, requiring no manual intervention.

[0067] Step 5: Operation End and Reset. Upon reaching the preset cutting length or detecting a cut-through signal (such as a sudden drop in load current), the main controller stops the cutting tool's rotation or shuts off the jet, and raises it to a safe height. Subsequently, the parallel robotic arm's linear travel module returns to its initial position, awaiting the next operation command. Throughout the entire process, the system log module records detailed sensor data and control commands at every moment, facilitating subsequent process optimization and fault tracing.

[0068] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A collaborative control system for high-precision heavy disassembly, characterized in that, The collaborative control system includes: a main controller, a parallel robotic arm linear motion module, a contact-type vertical tool setting module, and a laser ranging module; The parallel robotic arm linear travel module is used to drive the cutting execution end to perform high-rigidity linear feed along a preset dismantling path; The contact-type vertical tool setting module is located at the cutting execution end and is used to obtain vertical reference data of the surface of the object to be disassembled through physical contact before the disassembly operation. The laser ranging module is located at the cutting execution end and is used to monitor the distance data between the cutting execution end and the surface of the object to be disassembled in real time. The main controller is configured to calibrate the initial attitude based on the vertical reference data and correct the feed depth of the parallel robotic arm linear travel module in real time according to the distance data, so as to coordinate the control of the disassembly operation.

2. The collaborative control system for high-precision heavy disassembly according to claim 1, characterized in that, The parallel robotic arm linear motion module adopts a spatial parallel closed-loop linkage mechanism. The main controller controls the extension and retraction of each branch of the spatial parallel closed-loop linkage mechanism through inverse kinematics calculation, and uses the structural stiffness of the parallel mechanism to offset the reaction force during the heavy disassembly process.

3. The collaborative control system for high-precision heavy disassembly according to claim 1, characterized in that, The contact-type vertical tool setting module is equipped with a pressure-sensing probe. The vertical calibration logic of the main controller is as follows: control the parallel robotic arm linear walking module to drive the pressure sensing probe to slowly feed towards the surface of the object to be disassembled, and when a contact signal is received, lock the current pose coordinates as the vertical tool setting zero point.

4. The collaborative control system for high-precision heavy disassembly according to claim 1, characterized in that, The main controller is equipped with a dynamic depth compensation module, which compares the real-time distance data fed back by the laser ranging module with the preset cutting depth. When surface undulations are detected, the compensation amount of the parallel robotic arm linear travel module in the normal direction is adjusted in real time.

5. The collaborative control system for high-precision heavy disassembly according to claim 1, characterized in that, It also includes an environmental sensing and dust removal control unit, which is used to control the dust removal hood covering the cutting area to activate the negative pressure dust suction function according to the working status of the parallel robotic arm linear walking module.

6. The collaborative control system for high-precision heavy disassembly according to claim 1, characterized in that, The data interaction between the parallel robotic arm linear walking module, the contact vertical tool setting module, and the laser ranging module is all converged to the main controller via an industrial bus. The main controller is equipped with an adaptive motion algorithm for unstructured outdoor environments.

7. The collaborative control system for high-precision heavy disassembly according to claim 3, characterized in that, The main controller is also configured with multi-point fitting logic, which controls the contact vertical tool setting module to collect the coordinates of at least three non-collinear points on the surface of the object to be disassembled, fits the normal vector of the cutting plane, and corrects the entry angle of the cutting tool.

8. The collaborative control system for high-precision heavy disassembly according to claim 5, characterized in that, The main controller also integrates a sensor confidence dynamic reconfiguration module based on load gradient, which is configured to execute the following control logic: The rate of change of the driving current of the parallel robotic arm linear travel module is monitored in real time, and the rate of change of the driving current is used as a feedforward parameter to determine the intensity of dust outbreak in the cutting area. When the rate of change of the driving current exceeds a preset threshold, the main controller determines that it is a high dust and strong interference condition. At the same time, it triggers the environmental perception and dust removal control unit to start the negative pressure dust collection function and automatically reduces the confidence weight of the distance data fed back by the laser ranging module in the closed-loop control. Meanwhile, the main controller increases the confidence weight of the inverse kinematics-derived depth data based on the parallel robotic arm linear walking module, and uses the inverse kinematics-derived depth data to temporarily replace the distance data to maintain the cutting path until the rate of change of the driving current falls back to the steady-state range, so as to avoid laser ranging distortion caused by dust scattering during heavy disassembly.

9. The collaborative control system for high-precision heavy disassembly according to claim 2, characterized in that, The main controller is also equipped with a macro-micro coordinated stress deadlock relief module, which is configured as follows: During the disassembly process, the load status of the cutting execution end is monitored in real time. When a sudden increase in load is detected and the feed rate is forced to drop to zero, it is determined that a clamping deadlock caused by material stress release has occurred. The main controller immediately sends a macro-axis locking command to the parallel robotic arm linear walking module to force it to maintain its current position in the linear walking direction without moving backward, so as to serve as a stable mechanical support base. Simultaneously, the main controller controls the spatial parallel closed-loop linkage mechanism to enter the high-frequency micro-motion mode, driving the cutting execution end to perform micro-amplitude high-frequency reciprocating vibration in the tangential plane perpendicular to the feed direction. The high-frequency response characteristics of the parallel mechanism are used to loosen the deadlock contact surface. After the load monitoring value decreases, the cutting execution end is controlled to perform a micro-retraction action.

10. A robot for high-precision heavy disassembly, characterized in that, It includes a tracked chassis, a sunken electrical cabinet, an integrated host computer, and a collaborative control system as described in any one of claims 1 to 9; The parallel robotic arm linear travel module is mounted on the tracked chassis and is used to perform operations without hydraulic support legs. The recessed electrical cabinet is embedded in the internal space of the tracked chassis, and the integrated host computer is directly mounted on the robot body for human-machine interaction and issuing control commands.

Images