Self-adaptive single-tool trimming device and method for edge bonding machine

Abstract

The invention relates to edge sealing equipment, and discloses a self-adaptive trimming device and method based on a multi-degree-of-freedom single tool. The device adopts a single bracket structure, and two linear motion units arranged along the feeding direction and the thickness direction of a plate drive a cutter to realize two-dimensional linkage. The device is matched with an edge detection sensor to obtain the edge position and dip angle change of a plate, a control system generates a continuous space trajectory based on detection data and performs interpolation control on two movement units, and a cutter sequentially completes front-end trimming, upper and lower edge trimming and rear-end trimming within a machining period. The control system can predict the tail switching opportunity according to the plate length and the feeding speed, track continuity is kept in the switching process, and therefore a traditional double-blade mechanism is replaced with a single-blade structure. The device is compact in structure, small in debugging amount, capable of keeping stable trimming quality under different feeding conditions and suitable for edge sealing machining scenes with high requirements for trimming precision and equipment integration degree.

Description

Technical Field

[0001] This invention relates to the field of woodworking machinery and equipment technology, and in particular to an adaptive single-blade trimming device and method for edge banding machines. Background Technology

[0002] After edge banding the board, the edge banding machine usually needs to trim the front, back, and top and bottom edges to ensure the edge banding tape is flush and smooth with the board surface. The industry commonly uses a dual-support, dual-blade structure, with separate front and rear trimming blades, driven by two independent mechanisms. While this structure is mature, it suffers from several long-standing problems: First, the dual-blade mechanism has many components, drives, and adjustment points, resulting in a large overall size, long assembly time, and increased maintenance workload. Second, the two sets of blades operate independently, requiring manual adjustment for timing and trajectory matching. When encountering board size fluctuations or changes in feed speed, the trimming amount can be inconsistent, leading to unstable finished product quality. Third, existing devices often rely on unidirectional following or fixed trajectory modes. When the blades encounter complex curves, board vibrations, or rapid switching between front and rear trimming, it's difficult to maintain continuous and stable relative movement, resulting in vibration marks or gaps at the front entry and rear exit points. Furthermore, due to structural limitations, traditional devices struggle to simultaneously perform comprehensive edge trimming of the front, rear, and top and bottom edges within a single mechanism. This results in low equipment integration, making it difficult to adapt to new edge-banding equipment that demands higher compactness and response speed. Therefore, the industry urgently needs a new device with a simpler structure, faster response, the ability to adjust the tool trajectory in real time according to the sheet material condition, and the capacity to complete multiple edge trimming tasks with a single-blade structure, in order to overcome the inherent limitations of existing technologies.

[0003] Therefore, existing technologies still need to be improved. Summary of the Invention

[0004] Given that the existing technologies still have significant limitations in terms of structural redundancy, weak trajectory control capabilities, and unstable front-end and back-end trimming connections, especially in new edge-sealing equipment that requires compact structures and rapid responses, this invention proposes a technical solution that can replace the traditional double blades with a single blade and achieve overall trimming of the front end, upper and lower edges, and back end through multi-degree-of-freedom linkage and continuous trajectory planning. This solution aims to address the problem that existing devices cannot complete the entire trimming process in a single mechanism and are not sufficiently adaptive to speed changes.

[0005] The technical solution of the present invention is as follows: The present invention provides an adaptive single-blade trimming device for an edge banding machine, comprising a frame, a single bracket disposed on the side of the frame, and a blade mounted on the single bracket. The cutting tool is driven by a multi-degree-of-freedom adjustment mechanism with two-dimensional linkage capability. The adjustment mechanism includes: The first linear motion unit is set along the feeding direction of the sheet metal; A second linear motion unit is positioned along the thickness direction of the sheet material; The device further includes: Sensor components used for real-time detection of the edge position, tilt angle, and rate of change of sheet metal; A trajectory planning and control system electrically connected to the first linear motion unit, the second linear motion unit, and the sensor assembly; wherein... The control system is configured to generate a continuous spatial target trajectory based on the edge detection data of the board material, including a front-end trimming trajectory, a stable segment trimming trajectory, and a rear-end trimming trajectory. The control system generates a transition curve between the front-end trimming and the back-end trimming through trajectory interpolation, enabling the two linear motion units to execute in a continuous time manner, thereby completing all the trimming tasks of a traditional double-blade system under a single-blade structure. The control system predicts the back-end switching time based on the plate length and feed speed, enabling the tool to perform continuous processing of front-end trimming, upper and lower edge trimming, and back-end trimming without changing the number of supports and the number of tools.

[0006] In one embodiment, the first linear motion unit is driven by a servo linear motor, the second linear motion unit is driven by a stepper motor or a servo motor, and the two motion units are controlled by independent drive controllers to improve the dynamic response capability during trajectory interpolation.

[0007] In one embodiment, the sensor assembly includes a distance sensor, an angle sensor, and a wheel position detector. The control system performs fusion calculations based on the data from the three sets of sensors to obtain the real-time offset, local curvature, and rate of change of the edge of the board.

[0008] In one embodiment, the trajectory planning and control system includes a trajectory generation module, a trajectory interpolation module, and a speed prediction module. The trajectory generation module is used to generate a first continuous spatial trajectory based on the geometry of the front end of the sheet metal. The trajectory interpolation module is used to correct the tool trajectory based on the real-time offset during the stable trimming stage. The speed prediction module is used to adjust the acceleration constraints of the trajectory based on the changes in the sheet metal feed speed.

[0009] In one embodiment, when the control system detects that the tail of the board is about to enter the trimming area, it calculates the optimal switching time for the rear trimming based on the board length, current speed, and acceleration change trend, and generates a second continuous spatial trajectory for completing the rear trimming.

[0010] In one embodiment, the maximum acceleration ratio of the first linear motion unit and the second linear motion unit of the multi-degree-of-freedom adjustment mechanism satisfies A1 ≥ 2·A2, to ensure the response speed of the tool during the front-end and rear-end switching phase.

[0011] In one embodiment, when the control system performs front-end trimming, it first drives the second linear motion unit to perform pre-contact depth compensation, and then the first linear motion unit enters the tracking state, thereby reducing front-end cutting vibration.

[0012] In one embodiment, when the control system performs rear-end trimming, it first reduces the feed depth of the second linear motion unit and then executes the tail cutting trajectory of the first linear motion unit, so that the tool maintains a stable cutting state the moment the plate leaves the template.

[0013] In one embodiment, the continuous spatial trajectory includes a three-dimensional curved surface projection trajectory covering the upper edge, lower edge, and rounded corner areas of the plate, enabling a single blade to complete the trimming of multiple parts within a single mechanism.

[0014] Another aspect of the present invention provides an edge-sealing and trimming method based on single-blade tracking, applied to the aforementioned adaptive single-blade trimming device, the method comprising: (1) Collect the feed speed and initial position signals of the board when it enters the edge banding machine, and start the sensor assembly; (2) Calculate the real-time offset, local curvature and rate of change of the plate edge based on the laser ranging, angle detection and wheel feedback data; (3) The trajectory planning and control system generates a continuous spatial trajectory including the front end, straight section and the rear end; (4) Control the first linear motion unit to execute the trajectory along the plate direction, and control the second linear motion unit to perform cutting depth compensation; (5) In the front-end region, the trajectory is triggered to enter the straight section based on the rate of curvature change and depth pre-compensation; (6) Perform dynamic lateral compensation and speed matching based on real-time edge tracking in straight sections; (7) Before the predicted end of the plate arrives, execute the end trimming trajectory and smoothly retract the tool according to the feed rate and cumulative processing time.

[0015] In summary, this invention, through the combination of a single-blade structure and a multi-degree-of-freedom adjustment mechanism, enables the blade to continuously trim the front, upper and lower edges, and rear end within the same mechanism. Furthermore, through edge detection, trajectory generation, and real-time interpolation control, the blade's motion trajectory can be adjusted in real-time according to the shape of the sheet material and feed conditions. Compared to traditional dual-blade solutions, this invention not only reduces structural and debugging steps but also avoids misalignment between the front and rear blades, resulting in a more continuous trimming process with less impact. Relying on predictive switching and continuous trajectory control, the device maintains stable processing quality even with speed fluctuations, slight sheet material offsets, or local curvature changes, thereby improving the overall responsiveness of the equipment and the consistency of the finished product. It is suitable for edge banding processing equipment with higher requirements for structural compactness, automation, and trimming accuracy.

[0016] Compared to traditional dual-blade trimming mechanisms, this invention reduces the number of blades and drive components in its structure. However, by relying on a multi-degree-of-freedom linkage mechanism and continuous trajectory planning control, it achieves the front-end, top and bottom edges, and rear-end trimming tasks that conventional dual-blade or even multi-blade systems require. Especially noteworthy is the smooth blade movement during the transition between front-end and rear-end trimming. By predicting the tail position of the sheet metal and generating a time-continuous transition trajectory, the blades maintain stable motion—a feat difficult to achieve with existing structures. In practical use, even with changes in feed speed, slight sheet metal offset, or localized edge undulations, the blades maintain a stable relative distance. The trimming transition section no longer exhibits the vibration marks, edge chipping, or residual protrusions common in traditional devices, resulting in significantly improved trimming consistency.

[0017] More importantly, by achieving continuous trimming through a single-blade structure, the overall response of the device becomes faster, the number of adjustment points decreases, the structural compactness increases, and the machining quality becomes more stable. This result of "simpler structure but better machining effect" cannot be obtained by simply replacing the tool or adding sensors, but comes from the combined effect of multi-degree-of-freedom mechanisms, real-time interpolation control, and predictive switching, which is a comprehensive effect that is difficult to predict with existing technologies. Attached Figure Description

[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 A three-dimensional schematic diagram of the overall structure of an adaptive single-blade trimming device for an edge banding machine provided by the present invention (in a waiting position); Figure 2 A partially enlarged view of the tool assembly of an adaptive single-blade trimming device for an edge banding machine provided by the present invention shows the trimming blade in the working state; Figure 3 A schematic diagram of the working process (tracking trimming) of an adaptive single-blade trimming device for an edge banding machine provided by the present invention. Figure 4The present invention provides a flowchart of the steps of an adaptive single-blade trimming method for an edge banding machine. Detailed Implementation

[0019] To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention is further described in detail below. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. The embodiments of the invention are described below in conjunction with the accompanying drawings.

[0020] One embodiment of the present invention provides an adaptive single-blade trimming device for an edge banding machine. Please refer to [link to relevant documentation]. Figures 1-3 The system includes a frame 1, a single support 2 disposed on the side of the frame 1, and a cutting tool 3 mounted on the single support 2. The cutting tool 3 is driven by a multi-degree-of-freedom adjustment mechanism with two-dimensional linkage capability. The adjustment mechanism includes: The first linear motion unit 4 is set along the feed direction of the plate 9; The second linear motion unit 5 is set along the thickness direction of the plate 9; The device further includes: Sensor components for real-time detection of the edge position, tilt angle and rate of change of the plate 9; A trajectory planning and control system electrically connected to the first linear motion unit 4, the second linear motion unit 5, and the sensor assembly; wherein... The control system is configured to generate a continuous spatial target trajectory based on edge detection data of the board 9, including a front trimming trajectory 7, a stable segment trimming trajectory and a rear trimming trajectory 8. The control system generates a transition curve between the front-end trimming and the back-end trimming through trajectory interpolation, enabling the two linear motion units to execute in a continuous time manner, thereby completing all the trimming tasks of a traditional double-blade system under a single-blade structure. The control system predicts the back-end switching time based on the length of the plate 9 and the feed speed, so that the tool 3 can continuously perform front-end trimming, upper and lower edge trimming and back-end trimming without changing the number of supports and the number of tools 3.

[0021] Specifically, in this embodiment, the adaptive single-blade trimming device is installed on the side of the edge banding machine frame 1, and a single bracket 2 supports the blade 3 and its drive mechanism. The single bracket 2 is a rigid metal structure (e.g., welded aluminum alloy or steel), bolted to the frame 1 via a base, and equipped with anti-vibration pads to reduce vibration transmission. The single bracket 2 is equipped with a blade 3 mounting plate, a first linear motion unit 4 (X-axis), and a second linear motion unit 5 (Y-axis). The blade 3 is fixed along the mounting plate, and the shank and blade head of the blade 3 form a reliable rigid connection. The blade head is replaceable to adapt to different edge banding materials. Adjustment bolts and positioning pins are pre-installed on the bracket for on-site fine-tuning of the angle and height after factory testing.

[0022] The first linear motion unit 4 employs a servo-driven linear motor or servo slide (e.g., a combination of a ball screw or linear guide rail and a servo motor). Its stroke is selected based on the common length of the sheet metal 9 and the device structure, generally set to 100–300 mm (adjustable according to the model). The servo controller supports closed-loop control of position / speed / acceleration, with a position accuracy of ±0.05 mm and a maximum operating speed of 1.0 m / s (depending on the model). The first linear motion unit 4 is connected to the tool 3 mounting plate via a rigid coupling to ensure the relative position of the tool 3 remains stable during high-speed tracking. The driver supports external servo commands and a local interpolation interface, facilitating integration with the trajectory planning and control system.

[0023] The second linear motion unit 5 is used for the feed depth and upper / lower edge switching of the tool 3. It is recommended to use a precision stepper motor or a small-size servo motor with a ball screw or miniature linear guide. The stroke range is generally 10–30 mm to cover the thickness compensation and chamfering allowance of common edge banding materials. The second unit should be equipped with torque limiters or electronic torque monitoring to prevent overload. To ensure coordinated control, the driver of the second unit should support high-speed response (maximum acceleration recommended ≥ 2 m / s²) and be able to share a synchronous clock or bus (e.g., EtherCAT / CANopen) with the first unit to achieve real-time interpolation control.

[0024] The sensor assembly includes a set of non-contact distance sensors (laser or industrial ultrasonic / laser ranging modules), one or more angle sensors (such as small photoelectric encoders or magnetic angle sensors for tilt detection), and wheel position detection (wheel encoders or limit switches). The distance sensors are positioned 30–80 mm in front of the cutter 3 to detect the edge of the plate in advance; the angle sensors detect the angular deviation of the edge of the plate 9 relative to the cutter 3; and the wheel position detection confirms the relative position of the template and the plate 9. The sensor signals are sampled and filtered (e.g., 1–5 kHz sampling rate, low-pass filtered 50–200 Hz) and then subjected to simple Kalman filtering or weighted moving average within the controller to obtain smooth and low-latency edge position, tilt, and rate of change signals.

[0025] The trajectory planning and control system consists of a main control unit (industrial-grade PLC or embedded real-time controller) and a motion control module. The main control unit operates a trajectory generation module, a trajectory interpolation module, and a prediction switching module. The trajectory generation module generates a continuous spatial trajectory for the front end, stable segment, and rear end based on the feed start point, material geometry, and initial edge detection values ​​(using cubic spline curves or fifth-order polynomial interpolation to achieve curve smoothing). The trajectory interpolation module sends position / velocity commands to the servo driver at a period of at least 1 kHz to ensure continuous linkage between the two motion units. The control system supports communication with the edge banding machine's main control unit (such as MODBUS / Profinet) to obtain operating data such as feed speed and material length.

[0026] To avoid abrupt changes during the switching between the front and rear ends, the machining trajectory is connected by a curved transition section between the end of the front trajectory and the start of the rear trajectory. The transition section uses a three-dimensional continuous curve (continuous in position and first derivative). The interpolation algorithm ensures continuous velocity at the interpolation points, and acceleration limits are added when necessary (e.g., not exceeding the driver's allowable acceleration). The control system applies feedforward + feedback control at the trajectory transition: the feedforward is used to track the desired velocity distribution of the target trajectory, and the feedback loop uses PID control of position and velocity errors, and performs real-time fine-tuning of the interpolation path based on sensor feedback to ensure that tool 3 maintains a stable posture during the transition section.

[0027] Based on the known length of the sheet metal 9 or the measured entry point and feed rate, the control system calculates the expected tail-end arrival time and sets a preparation window before switching (e.g., 30–200 ms in advance, calculated based on the feed rate). Within the preparation window, the control system first gradually reduces the feed depth of the second linear motion unit 5 according to the set strategy (with a transition ratio of 10%–50%), while adjusting the speed distribution of the first unit to smooth the tail-end cutting. When the sensor feedback indicates that the tail-end of the sheet metal 9 has reached a critical point (the guide wheel signal or distance sensor reading reaches a threshold), the rear-end trimming trajectory 8 is executed, and the tool 3 is safely retracted to the standby position the instant the sheet metal 9 leaves, ensuring no edge chipping or residual protrusions. This logic can be calibrated using a parameter table at different feed rates.

[0028] During the stable tracking phase, the control system continuously calculates the offset and local curvature of the edge of the sheet metal 9. When the offset or curvature exceeds a preset threshold (e.g., offset > 0.2 mm or curvature radius < 5 mm), adaptive compensation is triggered: the tracking position of the first unit is adjusted and the depth of the second unit is corrected accordingly. If an action is detected that may cause a conflict or exceed the travel limit, the system enters a safety mode, issues an alarm, and handles the situation according to a preset strategy (slowing down the speed or temporarily lifting the tool). The system should record key operating conditions for post-event analysis and parameter optimization.

[0029] At the factory, the first and second linear motion units 5 are calibrated (zero position, velocity, acceleration, crosstalk coefficient), and edge curves of several standard plates 9 and trimming parameters of the cutter 3 are collected through teaching or automatic calibration programs to generate a default parameter library. On-site maintenance includes periodic checks of cutter 3 wear, guide rail lubrication, cleaning of sensor optical windows, and driver temperature monitoring. The device is equipped with convenient cutter head replacement and emergency stop devices to ensure production safety and ease of maintenance.

[0030] In a further embodiment, the first linear motion unit 4 is driven by a servo linear motor, and the second linear motion unit 5 is driven by a stepper motor or a servo motor. The two motion units are controlled by independent drive controllers to improve the dynamic response capability during trajectory interpolation.

[0031] Specifically, in this embodiment, the tool 3 mounted on the tool 3 mounting plate adopts a replaceable tool head structure. The tool head and the tool holder are fixed by a conical fit and screw locking, so that the tool 3 can maintain sufficient rigidity and vibration resistance during high-speed trimming. The installation angle of the tool 3 is precisely set at the factory through the scale adjustment area on the mounting plate. When changing the tool head, the operator only needs to readjust according to the scale to ensure the consistency of the cutting angle and avoid angle drift caused by tool replacement. The tool 3 mounting plate is fixed to the first linear motion unit 4 using an integral connecting seat. This connecting seat is made of high-rigidity aluminum alloy and has been anodized to prevent the plate surface from bending or twisting when following the motion at large acceleration. The underside of the tool 3 mounting plate is also provided with vibration damping pads to further improve the contact stability between the blade and the edge of the plate 9 under high-frequency micro-vibration. In this embodiment, the safety cover of the tool 3 is fixed by quick-release buckles. Its cover has a long strip opening, which allows the operator to observe the wear of the tool 3 without stopping the machine, while not affecting the field of view of the front sensor. The entire installation structure remains stable during frequent switching between upper and lower edges, ensuring that the position of the cutter 3 relative to the edge of the plate 9 can move precisely according to the instructions of the trajectory planning system, thereby achieving continuous trimming of the front end, straight section and rear end.

[0032] To accommodate edge banding strips of varying thicknesses and materials, the cutting tool 3 in this embodiment can be fitted with different blade shapes, such as R-angle tools, 45° chamfering tools, or straight-blade tools, depending on the processing task. The tool 3 mounting plate has multiple pre-drilled mounting holes, allowing tools 3 of different lengths or diameters to be installed within a reasonable overhang distance, preventing cutting chatter caused by excessive tool overhang. During the debugging phase, the cutting depth is calibrated by fine-tuning the tool 3's extension using the second linear motion unit 5. The controller stores this calibration value as a parameter to maintain a consistent trimming allowance under different sheet metal and speed conditions. If the cutting force increases due to wear during operation, and the sensor detects abnormal vibration or changes in the force on the guide roller, an alarm will be automatically triggered to remind the operator to replace the tool head. The entire tool 3 mounting and cutting structure maintains trimming accuracy without drifting even after prolonged operation, thanks to the structural rigidity of the mounting plate and the tool holder fixing method.

[0033] In a further embodiment, the sensor assembly includes a distance sensor, an angle sensor, and a wheel position detector. The control system performs fusion calculations based on the three sets of sensor data to obtain the real-time offset, local curvature, and rate of change of the edge of the plate 9.

[0034] Specifically, the edge detection sensor assembly in this embodiment includes a laser rangefinder located in front of the cutter 3, and a set of angle sensors for determining the wheel's attitude. The laser rangefinder is installed approximately 50 mm in front of the cutter 3, enabling it to detect edge position changes before the edge of the sheet metal 9 enters the cutting zone. The sensor is fixed on a separate small sensor bracket, which is isolated from the main bracket by anti-vibration rubber pads, effectively reducing the interference of cutting vibration on the rangefinder signal. The sampling frequency of the laser sensor is set to 2–5 kHz and undergoes first- or second-order low-pass filtering in the controller, ensuring that the signal maintains a fast response time while suppressing instantaneous jumps caused by local textures in the sheet metal 9. The wheel angle sensors are used to determine whether there is slight warping or angular deviation at the edge of the sheet metal 9. This data is combined with the rangefinder information in the control system to obtain a more stable edge profile.

[0035] When there are slight fluctuations, local gaps, or uneven adhesive lines on the edge of the sheet 9, the sensor system can reflect these changes to the trajectory planning and control system in real time. After collecting the changes in the ranging value, the controller will make minor corrections to the target trajectory based on the offset, the rate of change, and the feed speed of the sheet 9, so that the tool 3 can maintain a constant relative distance with the edge. Especially in the stages of entering the front end of the sheet 9 and leaving the rear end of the sheet 9, the wheel detection and laser ranging data need to be judged in coordination to avoid the tool 3 from accidentally rushing or delaying its return due to the sudden disappearance of the edge. In this embodiment, through the fusion processing of sensor data, the control system can maintain a stable cutting posture under complex edge conditions, effectively reducing the edge chipping, slight gaps, or overcutting phenomena commonly found in traditional devices. Since the delay of the sensor data processing module is controlled within 1–2 ms, the time of the entire detection-response link fully meets the requirements of high-speed edge banding processing.

[0036] In a further embodiment, the trajectory planning and control system includes a trajectory generation module, a trajectory interpolation module, and a speed prediction module. The trajectory generation module is used to generate a first continuous spatial trajectory based on the geometry of the front end of the sheet metal 9. The trajectory interpolation module is used to correct the trajectory of the tool 3 based on the real-time offset during the stable trimming stage. The speed prediction module is used to adjust the acceleration constraint conditions of the trajectory based on the change in the feed speed of the sheet metal 9.

[0037] Specifically, the trajectory planning and control system described in this embodiment consists of three cooperative functional modules: a trajectory generation module, a trajectory interpolation module, and a speed prediction module. Its hardware implementation can be achieved by combining an industrial real-time controller (such as an embedded real-time controller or industrial PLC that supports EtherCAT) with a dedicated motion control card. The controller communicates with the servo drivers of the first linear motion unit 4 and the second linear motion unit 5 through a high-speed fieldbus, and maintains data communication with the main control of the edge banding machine (feed speed, board 9 arrival signal).

[0038] This architecture ensures deterministic data transmission while facilitating subsequent software upgrades and on-site parameter adjustments.

[0039] After the plate 9 enters the detection area and obtains initial edge information, the trajectory generation module generates a smooth three-dimensional spatial reference trajectory based on the geometric features of the front end of the plate 9, the preset trimming allowance, and the geometric parameters of the tool 3. This trajectory is generated using a cubic spline or fifth-order polynomial interpolation algorithm. The trajectory is spatially segmented into front, stable, and rear segments, but the position and first derivative are ensured to be continuous at the segment connections. During generation, the safe distance between the tool 3 and the guide wheel and the cutting width of the tool 3 are considered, and a speed distribution suggestion (i.e., the target speed level for each segment) is provided to the trajectory interpolation module to balance machining quality and equipment dynamic response capabilities. The trajectory interpolation module operates at a fixed cycle (1 kHz or higher is recommended, depending on the servo system capability), discretizing the continuous reference trajectory from the trajectory generation module into a sequence of position / velocity / acceleration commands acceptable to the driver, and implementing a feedforward-feedback control strategy. The feedforward part issues the desired position and velocity according to the planned speed distribution, while the feedback part collects the real-time position / velocity errors from the servo encoder and sensors, using a cascaded structure of a position loop PID and a speed loop PI for error compensation. Simultaneously, acceleration saturation and chatter suppression filtering are incorporated into the interpolation calculation (e.g., limiting the rate of change of velocity at the interpolation point and using low-pass filtering in the position loop to remove high-frequency noise) to ensure that no abrupt changes or resonance occur when the two motion units are linked. The interpolation module also supports an insert-type fine-tuning interface, allowing sub-pixel-level corrections to the current interpolation point based on sensor fusion results, thereby maintaining a constant relative position between the tool 3 and the edge of the board under real-time conditions. The speed prediction module continuously receives feed speed and acceleration information from the edge banding machine's main control unit and, combined with the known length of the board 9 or the entry timing measured by the sensor, calculates the optimal trigger window for end-point switching in real time.

[0040] This module employs a combination of a simple physical prediction model (linear or second-order prediction model) and a sliding window averaging method to reduce judgment errors caused by single-frame anomalies. A confidence assessment is applied to the prediction results; an automatic switching scheme is triggered only when the confidence level exceeds a threshold, thus avoiding unnecessary actions due to misjudgments. The switching time obtained from the velocity prediction is simultaneously used as input to the trajectory interpolation module, allowing the interpolation module to smoothly transition according to a predetermined curve in advance, thereby achieving time-continuous linkage actions. To improve robustness under complex edge conditions, the trajectory planning and control system implements sensor data fusion and online soft-limiting functions. Sensor fusion uses weighted Kalman filtering or complementary filtering to fuse data from laser ranging, angle sensors, and wheel encoders into smooth edge position and local curvature estimates. Based on this estimate, the trajectory generation or interpolation module can perform small trajectory corrections (e.g., position correction within ±0.2 mm) within milliseconds and issue the corrected interpolation commands in real time. If the fusion results show a sudden change exceeding the safety threshold (such as instantaneous offset >1 mm or drastic inconsistency in sensor data), the system will prioritize a safety strategy: first, reduce the speed of tool 3 and lift the tool; then, resume processing only after confirming the edge has disappeared or stabilized, ensuring the safety of the equipment and workpiece. The control system software design includes a parameter library and an on-site parameter adjustment interface. When generating the trajectory, preset curve templates for different materials and edge banding types can be called, and template parameters can be updated through teaching or offline optimization. A set of default parameters (tool 3 geometry, trimming allowance, speed level, acceleration limit) is provided as standard. On-site commissioning personnel can fine-tune parameters such as the speed prediction window, interpolation cycle, and filter bandwidth according to the actual production line conditions. The controller supports operating condition log recording and playback functions, facilitating the analysis of abnormal operating conditions and continuous improvement of the control strategy.

[0041] In a further embodiment, when the control system detects that the tail of the board 9 is about to enter the trimming area, it calculates the optimal switching time for the rear trimming based on the length of the board 9, the current speed, and the acceleration change trend, and generates a second continuous spatial trajectory for completing the rear trimming.

[0042] More specifically, in this embodiment, the switching of the tool 3 between the front end, the stable section, and the rear end is not triggered by a traditional fixed point, but is automatically completed by the trajectory planning and control system based on the real-time feed speed of the sheet metal 9, edge sensor feedback, and the current posture of the tool 3. The determination of the end of front-end cutting is mainly based on the distance curve change of the laser rangefinder. When the sensor continuously collects the process of the edge of the sheet metal 9 moving from approaching to stabilizing, the controller will determine whether the front end has completely entered the stable cutting zone by the rate of curvature change. If the rate of change is below the threshold for several consecutive cycles, the system considers the front-end cutting to be completed, and then the trajectory automatically switches to the straight section following mode. In the determination of rear-end cutting, the controller calculates the theoretical arrival time of the tail end based on the feed speed and entry time recorded when the sheet metal 9 enters, and combines the edge rapid retreat signal detected by the sensor (such as a sudden increase in the distance value or a sudden change in the wheel rotation angle) to determine the actual tail end position, so that the tool 3 can perform rear end trimming and safely retract the tool at the most appropriate time. Throughout the switching process, the controller always maintains the continuity of the command speed, and there will be no edge chipping or residual protrusions caused by early or delayed switching as in traditional devices.

[0043] To improve the system's adaptability to different materials and feed rates, this embodiment allows operators to set a switching sensitivity parameter in the debugging interface. This parameter controls the threshold size for front-end and back-end switching decisions. For example, when machining harder MDF, the sensitivity can be lowered to prioritize system stability; while when machining softer multilayer boards, the sensitivity can be increased to prevent slight overcutting by tool 3 on flexible materials. This parameter simultaneously affects the rate of change threshold for the end-of-stage decision, the size of the back-end prediction window, and the speed convergence strategy of the interpolation trajectory, making the entire switching action more consistent with the characteristics of on-site production. The final result is that tool 3 basically does not produce sudden jumps at the switching nodes, especially maintaining good consistency of cutting marks at high-speed feeds.

[0044] In a further embodiment, the maximum acceleration ratio of the first linear motion unit 4 and the second linear motion unit 5 of the multi-degree-of-freedom adjustment mechanism satisfies A1≥2·A2, so as to ensure the response speed of the tool 3 during the front-end and rear-end switching phase.

[0045] Specifically, the real-time edge tracking function in this embodiment relies on the coordinated fusion of multi-sensor data and high-speed interpolation. When the sheet metal 9 enters the stable cutting section, the laser rangefinder continuously outputs the edge position, while the wheel angle sensor provides the local edge tilt angle. These two types of data are weighted and fused within the controller to obtain the target lateral compensation amount required by the current cutting tool 3. The controller adds this compensation amount to the trajectory interpolation point at a frequency above 1 kHz, enabling the first linear motion unit 4 to adjust its position in real time to match the edge shape. To avoid momentary misjudgments caused by local gaps in the edge banding or abnormal wood grain reflection, the controller performs short-window smoothing on the sensor data and employs a "suppression strategy" when a sudden change is detected: first, maintaining the compensation direction from the previous moment, and then judging whether adjustment is indeed necessary based on the trend of the next two cycles, thereby ensuring that the tracking action is both sensitive and not excessively affected by noise.

[0046] When slight warping or a sudden increase in edge curvature occurs on the sheet metal 9, the control system can quickly detect the anomaly through changes in angle sensor readings and instruct the second linear motion unit 5 to slightly adjust the longitudinal extension of the tool 3, thus stabilizing the contact depth of the tool 3 relative to the edge. For example, when machining curved edges or areas with local unevenness, the system can automatically adjust the feed depth according to the edge curvature, keeping the cutting load within a safe range and reducing tool marks or chipping caused by sudden load increases. The delay of the entire real-time tracking action is controlled within a few milliseconds. Combined with the fast response of the servo system, the tool 3 trajectory can maintain a smooth and reliable following effect even at high-speed feeds (e.g., 15–25 m / min).

[0047] In a further embodiment, when the control system performs front-end trimming, it first drives the second linear motion unit 5 to perform pre-contact depth compensation, and then the first linear motion unit 4 enters the tracking state, thereby reducing front-end cutting vibration.

[0048] To further improve processing stability, this embodiment introduces edge curvature estimation and adaptive smoothing algorithms based on real-time tracking. When the controller detects a significant change in the edge curvature of a certain area of ​​the sheet metal 9, it automatically inserts a micro-buffer zone into that trajectory segment. This generates a fine-tuning space around the original trajectory, allowing the first linear motion unit 4 to adjust in segments over shorter distances, rather than completing a large displacement all at once. This avoids the tool 3 producing obvious cutting marks at points of drastic curvature change. The generation of this buffer zone is automatically completed based on sensor fusion results, requiring no operator intervention. The system automatically determines whether to activate the buffer strategy based on a preset curvature threshold. In this way, even if the sheet metal 9 itself has slight geometric errors or local fluctuations in the edge banding, the tool 3 maintains a stable posture and uniform force during tracking.

[0049] When encountering materials with frequent edge variations and high noise levels (such as particleboard or decorative panels with obvious texture reflections), the system automatically increases the smoothing coefficient and reduces the high-frequency compensation amplitude through algorithms. This makes the cutter 3 move more smoothly, preventing excessive wear or trajectory jitter caused by frequent micro-tracking. Conversely, on materials with relatively straight edges and stable textures, the system automatically reduces the smoothing coefficient and increases the response speed, allowing the cutter 3 to more closely conform to the real edge. This amplitude adjustment process is entirely based on the statistical characteristics of real-time sensor data, such as variance, curvature gradient, and short-term stability indicators, which is a typical adaptive control strategy. The final result is that the system can maintain ideal trimming consistency under various material conditions, without significant tool mark fluctuations caused by environmental or material differences.

[0050] In a further embodiment, when the control system performs the back-end trimming, it first reduces the feed depth of the second linear motion unit 5, and then executes the tail cutting trajectory of the first linear motion unit 4, so that the tool 3 maintains a stable cutting state at the moment the plate 9 leaves the template.

[0051] In this embodiment, the tool 3 employs a real-time speed matching strategy when following the edge of the sheet 9. The trajectory planning and control system continuously receives the feed speed signal from the edge banding machine's main control unit, and after limiting and smoothing this speed, uses it as the target speed for the tool 3 along the direction of the first linear motion unit 4. This ensures that the tool 3 always follows the sheet 9 at the same speed. To prevent slight speed fluctuations caused by local friction differences in the main conveyor roller of the edge banding machine from causing vibration in the tool 3, the system adds a short-window filter to the speed signal and limits the rate of speed change to within the equipment's tolerance range. For example, the acceleration change within 10 ms does not exceed the system's set safety threshold, thus making the tool 3's movement smoother. During this process, the tool 3 maintains a constant relative position to the edge of the sheet 9, and the second linear motion unit 5 only makes minor adjustments when there are local undulations or angle changes at the edge. This results in significantly better processing quality in the straight section compared to traditional mechanical following structures. Because the following speed is completely synchronized with the feed speed of the plate 9, the tool 3 will not leave fine ripples on the surface of the plate 9 due to delay or lag. Especially under high-speed feed conditions (e.g., above 20m / min), it can still maintain consistent cutting marks and neat edge banding cross-section.

[0052] To further improve the stability of long board processing, the system also implements a dynamic anti-drift mechanism in the straight section. When the sensor detects a slow drift trend in the board 9 caused by uneven pressure from the pressure rollers or slight bending of the board 9 itself, the real-time edge tracking module provides a small compensation amount. However, this compensation amount is added to the trajectory interpolation point with a smaller coefficient in the straight section, so that the tool 3 does not immediately react violently to occasional small deviations, but gradually recovers to the target position in a manner similar to "buffered following". From the on-site processing results, this buffering mechanism can significantly reduce the multiple slight fluctuations caused by the uneven material of the board 9, making the processed edge banding more linear and smooth.

[0053] In a further embodiment, the continuous spatial trajectory includes a three-dimensional curved surface projection trajectory covering the upper edge, lower edge, and rounded corner areas of the plate 9, enabling a single blade to complete the trimming of multiple parts within a single mechanism.

[0054] Specifically, to ensure the continuity of the tool 3's switching process between the front, straight, and rear trajectories, the trajectory planning control system in this embodiment reserves transition buffer sections at the joints of each segment when generating the trajectory. The buffer sections are generated based on cubic spline functions, ensuring that the trajectory is not only positionally continuous at the joints but also that the first derivative remains continuous. This design guarantees that the tool 3 will not experience sudden speed changes during the switching segments, and the servo system will not need to withstand large acceleration impacts, thereby extending the lifespan of the mechanical structure and the tool 3. In actual operation, the system automatically fine-tunes the length of the buffer sections based on the material of the edge banding and the edge shape of the board 9. For example, when processing hardwood edge banding, the buffer section is appropriately extended to reduce tool marks caused by sudden changes in cutting load; when processing flexible edge banding, the buffer section is appropriately shortened to improve overall efficiency.

[0055] In terms of control strategy, the interpolation algorithm of the buffer section adopts a combination of feedforward and feedback. The feedforward end outputs the target position based on the speed distribution set during trajectory generation, while the feedback end corrects the interpolation point based on the encoder data of the servo driver and the sensor feedback value, so that the tool 3 keeps close to the real edge during high dynamic processes. When switching from the front end to the flat section, the system automatically performs gradual compensation based on the difference between the current real-time speed of the tool 3 and the target speed, so that the speed of the tool 3 smoothly converges to the speed of the flat section within tens of milliseconds. When entering the rear end cutting, the system also adopts a similar gradual strategy, so that the tool 3 gradually retracts to the safe return path within a limited distance, rather than suddenly lifting the tool, thereby avoiding leaving obvious cutting marks on the tail end of the plate 9. This continuous switching mechanism shows high stability in actual operation, so that operators do not need to frequently adjust when switching between different materials and thicknesses.

[0056] In a further embodiment, the trajectory planning and control system adaptively adjusts the speed distribution of the first linear motion unit 4 according to the local curvature of the edge of the plate 9, thereby maintaining a uniform trimming amount under complex contour conditions.

[0057] During the tail-end processing stage of sheet metal 9, the speed prediction module in this embodiment calculates the expected time when the tail end will reach the position of tool 3 based on the initial position signal of sheet metal 9 when it enters the machine, the real-time feed rate, and the cumulative processing time, and generates an advance cutting buffer. Within this buffer, the controller reduces the weight of edge tracking in the trajectory interpolation and gradually guides tool 3 into the preset tail-end trimming trajectory, making the tool 3 posture smoother during the tail-end transition. If the sensor detects a sudden retreat of the edge, such as a continuous increase in the laser ranging value within a very short time, the controller determines that sheet metal 9 is about to leave the cutting area and immediately performs a tool retraction action according to the tail-end trajectory to ensure that tool 3 does not overcut at the tail-end position.

[0058] To compensate for the uncertainty in tail-end processing under different speed conditions, the system records the deviation between the tail-end response time and the sensor trigger time after each processing step. Based on this deviation, it automatically fine-tunes the parameters of the prediction model, making the tail-end calculation increasingly closer to actual working conditions. This online self-learning method makes the device more stable after long-term use, maintaining high consistency even when switching between different plate lengths and feed speeds in the factory. Furthermore, in abnormal situations, such as a sudden deceleration of plate 9 due to external force or friction changes, the system can promptly identify this trend within the prediction window and automatically extend the buffer time, thereby avoiding the risk of incorrect tail-end cutting or delayed tool retraction. The final result is that regardless of whether the feed speed is stable, tool 3 can achieve a natural, clean, and scratch-free cutting effect during tail-end processing.

[0059] Another embodiment of the present invention provides an edge-sealing and trimming method based on single-blade tracking, applied to the aforementioned adaptive single-blade trimming device. Please refer to [link to relevant documentation]. Figure 4 The method includes: S1. Collect the feed speed and initial position signal of the board 9 when it enters the edge banding machine, and start the sensor assembly; When sheet material 9 enters the detection area, the controller first receives the feed speed and feed start time signals from the edge banding machine's main control, and simultaneously wakes up the edge detection sensor components to acquire data. The system sets the initial sampling frequency (1–5 kHz recommended, 2 kHz commonly used) for the laser rangefinder, angle sensor, and guide wheel encoder, and performs a short self-check during the hard start phase: checking for sensor over-limit values, signal loss, or significant offset; if the self-check passes, the "initial position signal" and feed speed of this processing are written into the running buffer as the time reference for subsequent prediction and trajectory generation; if the self-check fails, the system enters safe mode and prompts for manual inspection. This step ensures the reliability of the sensor data and provides a unified reference for time and position in subsequent steps.

[0060] In a further embodiment, the method further includes: S2. Calculate the real-time offset, local curvature, and rate of change of the edge of the plate 9 based on the laser ranging, angle detection, and feedback data from the guide wheel; Specifically, sensor data is filtered through a short window (e.g., a 1–5 period moving average or a first-order low-pass filter) and then fused within the controller to calculate real-time offset, local curvature, and rate of change. In practice, the laser ranging output is time-aligned (based on the wheel encoder or master clock), and then weighted Kalman filtering or complementary filtering is used to fuse angle and distance data to obtain the optimal estimate of the edge of plate 9. Curvature estimation can be approximated by performing a quadratic polynomial fitting on several local points to obtain the second derivative, while the rate of change is the first-order time difference of the offset. The system simultaneously calculates signal confidence (based on sensor consistency and noise variance). When the confidence is below a threshold (e.g., <0.6), a fault-tolerant strategy is triggered: reducing the speed tracking weight or briefly raising the blade to avoid malfunctions.

[0061] In a further embodiment, the method further includes: S3. The trajectory planning and control system generates a continuous spatial trajectory including the front end, straight section and the rear end. The trajectory planning and control system, based on the fused edge estimation, combines tool 3 geometric parameters (tool width, tool tip offset), preset trimming allowance, and feed rate samples to generate a continuous spatial trajectory covering the front, straight section, and rear end. Trajectory generation preferentially uses cubic spline or fifth-order polynomial interpolation to ensure position and first-order derivative continuity. The trajectory is represented spatially as a point sequence with a target velocity distribution. Safety constraints are considered during generation: the minimum safe distance between tool 3 and the template, the maximum stroke of the two motion units, and the upper limit of acceleration. The trajectory segments (front end / stable / rear end) logically exist but are numerically connected by smooth transition segments, preventing abrupt command changes during interpolation.

[0062] In a further embodiment, the method further includes: S4. Control the first linear motion unit 4 to execute the trajectory along the direction of the plate 9, and control the second linear motion unit 5 to perform cutting depth compensation. In interpolation execution, the controller discretizes the trajectory into position / velocity / acceleration commands acceptable to the driver (interpolation period recommended 1 kHz). The first linear motion unit 4 is responsible for following and velocity matching along the plate 9, while the second linear motion unit 5 performs depth compensation and fine-tuning of the upper and lower edges. Actual control employs feedforward-feedback coupling: the feedforward loop issues the desired trajectory at the target velocity generated by the trajectory, and the feedback loop uses position and velocity errors as inputs to perform PID / PI adjustments, and performs feedforward acceleration correction when necessary to suppress hysteresis. To reduce coupling errors, time synchronization commands are used between the two axes (interpolation points at the same timestamp are issued simultaneously), and an inter-axis coordinator is maintained within the controller to handle velocity distribution and acceleration limits.

[0063] In a further embodiment, the method further includes: S5. Trigger entry into the straight section trajectory in the front-end region based on the rate of curvature change and depth pre-compensation. The determination and transition of the front-end entry are based on the rate of curvature change and pre-compensation logic. When the rate of curvature change is lower than a set threshold and the distance measurement value tends to stabilize within a certain number of consecutive sampling periods, the system considers that the tool 3 has entered the straight section for sustainable tracking. Before this, the second linear motion unit 5 first performs "pre-contact depth compensation" according to the preset curvature compensation curve (e.g., first reaching 30%–70% of the target cutting depth) to establish a stable cutting load. Subsequently, the first linear motion unit 4 gradually enters the tracking speed. The interpolation at the transition point adopts a speed gradient (S-shaped transition) and limits the acceleration to not exceed the upper limit allowed by the driver, thereby minimizing front-end entry vibration and cut defects. The operating interface allows engineers to adjust the curvature threshold and pre-compensation ratio to match different materials.

[0064] In a further embodiment, the method further includes: S6. Perform dynamic lateral compensation and speed matching based on real-time edge tracking in straight sections; In the stable phase, the system uses the fused edge data after short-window filtering as the lateral compensation target, calculates the lateral correction amount for each interpolation point in real time, and sends it synchronously to both motion units. For speed matching, the first axis speed directly follows the feed speed of the edge banding machine's main control (after short-window smoothing and rate limiting), and speed distribution optimization is applied within the interpolation module to reduce energy consumption and chattering. Lateral compensation is preferentially performed using small-amplitude, high-frequency fine-tuning (e.g., compensation amount is typically within ±0.2 mm, delay <5 ms), and a steady-state strategy is adopted when encountering sudden anomalies (e.g., instantaneous offset >1 mm): first, the position is maintained according to the historical average, and then a larger compensation is performed after subsequent sampling confirms the trend of change, preventing noise-driven over-adjustment. During this stage, key operating condition data (deviation, compensation amount, tool torque) are recorded for subsequent analysis and online parameter optimization.

[0065] In a further embodiment, the method further includes: S7. Before the predicted arrival time of the tail end of plate 9, execute the tail end trimming trajectory and smoothly retract the tool according to the feed rate and cumulative processing time.

[0066] The tail-end processing employs a dual mechanism based on prediction and sensor confirmation. The prediction section estimates the expected arrival time window of the tail end using feed rate, plate entry time (9), and cumulative processing time. Within this window, it gradually reduces the cutting depth of the second axis and adjusts the speed distribution of the first axis to create a retraction buffer. The sensor confirmation section uses abrupt changes in the guide wheel status and distance measurement signal as trigger points. Once rapid edge retraction or guide wheel depressurization is detected, the controller immediately executes the preset tail-end trimming trajectory and triggers a smooth retraction action (the retraction path is a reverse interpolation along a calculated continuous trajectory, with limited retraction acceleration to avoid springback). After each processing cycle, the system records the deviation between the predicted arrival time and the actual trigger time in the model self-learning module for online correction of prediction parameters (e.g., adjusting the prediction window advance). As usage frequency increases, the timing of the tail-end processing becomes increasingly accurate, resulting in more natural and residue-free retraction. If a safety risk is detected at any step of the process (e.g., driver malfunction, travel limit triggering, or severe sensor inconsistency), the system will prioritize tool lifting and issue an alarm to ensure equipment and workpiece safety.

[0067] In a further embodiment, during the real-time edge tracking step, by fusing and short-window filtering the sensor data, lateral compensation is completed with a millisecond-level delay, maintaining stable contact between the tool 3 and the edge of the plate 9.

[0068] Specifically, in the real-time edge tracking stage, sensor data is first preprocessed at the acquisition end using a short window (e.g., moving average or first-order low-pass filtering for 1–5 sampling periods, sampling frequency 1–5 kHz, commonly 2 kHz), and then sent to the fusion module according to the timestamp and aligned with the wheel encoder. The fusion module uses a weighted Kalman filter or complementary filter framework. The initial weight values ​​can be set according to sensor reliability: 0.6 for laser ranging, 0.3 for angle sensors, and 0.1 for wheel encoders (adjustable on-site), and the weights are dynamically adjusted based on the signal-to-noise ratio during operation. The fusion result output includes edge lateral deviation, local curvature, and confidence index. The interpolation module calculates the lateral compensation amount based on this deviation and sends it to the two motion units within a 1 kHz (or higher) interpolation period. To control latency and avoid noise-driven over-action, the compensation employs a small-amplitude, high-frequency fine-tuning strategy with a dead zone (typical compensation ±0.2 mm, short-time suppression threshold 0.05–0.1 mm), and imposes speed and acceleration limits on the compensation output (e.g., maximum lateral speed 0.1–0.5 m / s, maximum lateral acceleration 2 m / s²). When the fusion confidence falls below a set threshold (e.g., 0.6) or a transient jitter (deviation abrupt change >1 mm) is detected, the system enters a brief fault-tolerant state: maintaining the last effective compensation, reducing the tracking weight, or lifting the tool according to the strategy, and resuming tracking after the data stabilizes. The time delay of all tracking actions (from sensor sampling to compensation action issuance) is controlled within a few milliseconds to ensure "millisecond-level" contact of the tool's three pairs of edges.

[0069] In a further embodiment, a continuous velocity and acceleration buffering strategy is adopted during the trajectory switching phase to achieve continuity of position, velocity and acceleration between the three trajectory segments: the front end, the straight section and the rear end.

[0070] Specifically, during the trajectory switching phase, trajectory generation is constructed in segments (front end / straight / rear end), but numerically, it is connected by continuous S-shaped or cubic spline transitions to ensure continuity of position, velocity, and acceleration. Specifically, a buffer zone (e.g., 5–50 mm, linearly scaling with feed rate) is reserved at the connection points of adjacent segments, using cubic splines or fifth-order polynomials to ensure continuity of the first and second derivatives at the connection points. The interpolator gradually changes the target velocity within these buffer zones (smoothing from the current velocity to the next target velocity segment), while acceleration and jerk are limited (e.g., jerk limited to 5–20 m / s³, specifically set according to the model and drive capability) to prevent sudden shocks to the drive unit and mechanical structure. A feedforward-feedback approach is used: the feedforward sends the desired position / velocity according to a smooth curve, and the feedback loop uses the position / velocity error for PI / PID correction and adds a low-pass vibration damping filter. If the buffer section detects an abnormal edge (short-term large deviation) indicated by the sensor, the interpolator will dynamically extend the buffer and reduce the speed to ensure a smooth transition; conversely, when the edge is stable and efficiency needs to be improved, the buffer length can be dynamically shortened. All switching strategies are saved as templates in the controller parameter library and can be switched with one click according to material type, edge banding type, and production line speed, ensuring both a seamless transition and production capacity.

[0071] In a further embodiment, the system automatically corrects the prediction model parameters online based on the tail trigger deviation after each machining operation to improve the stability and consistency of the tail retraction.

[0072] Specifically, regarding tail-end prediction and online correction, a dual strategy of "prediction-driven + confirmation-triggered" is adopted: the prediction module uses a simple physical model (linear or second-order equation of motion) based on feed rate, entry time, and cumulative machining time to predict the tail-end arrival time window and outputs a lead time (e.g., a lead time window of 30–200 ms based on speed). Within this lead time window, the controller gradually reduces the second-axis depth compensation, adjusts the first-axis speed distribution, and enters the retraction buffer state. Simultaneously, sensor confirmation (wheel pressure loss, sudden increase in distance measurement, etc.) serves as the final trigger signal; once confirmed, smooth retraction is immediately executed. After each machining operation, the system records the deviation between the predicted trigger time and the actual confirmed trigger time, and uses exponentially weighted moving average (EWMA) or small-step parameter adjustments (e.g., updating the lead time by 1–5% each time) to online correct the key parameters of the prediction model, enabling the model to adapt to production line characteristics during operation. If the deviation consistently exceeds the upper limit (e.g., >20% or absolute error >50 ms), the system marks the batch as abnormal and suggests manual inspection of the sensors or speed stability. All end-processing logic has a clear safety backoff mechanism: when the driver reports an error, the travel is in place, or the sensors are inconsistent, the tool is immediately lifted and an alarm is triggered to prevent accidental cutting or damage to the workpiece.

[0073] In summary, this invention addresses the dynamic uncertainty of single-blade trimming in the edge-sealing process, constructing a compact, coherent, and adaptive trimming system. From the installation of the tool 3 and the arrangement of the dual linear motion units, to the data fusion of multiple sensors, and then to trajectory planning, real-time tracking, and tail-end prediction, each step is designed to improve cutting stability, reduce switching vibration marks, and enhance processing consistency. In actual operation, the linkage between the modules is not a simple, independent superposition, but rather, through continuous trajectory, speed matching, and adaptive compensation, the tool 3 maintains a "natural and smooth" cutting state at different processing stages. Even if the sheet metal 9 has slight warping, uneven feed, or local edge changes, it can maintain a stable following posture.

[0074] From the perspective of on-site debugging, the solution of this invention reduces reliance on operator experience and eliminates the need for frequent fine-tuning of the tool position, achieving repeatable trimming quality across most materials and speed ranges. During long-term operation, the system can continuously optimize its tail-end prediction and compensation characteristics through parameter self-learning, making the entire mechanism increasingly "intelligent" with use. Therefore, this invention is not a minor modification to an existing structure, but rather a trimming platform built around the concept of adaptive control, possessing significant potential for widespread adoption and capable of stable implementation on various types of edge banding equipment.

[0075] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. An adaptive single-blade trimming device for an edge banding machine, comprising a frame, a single support disposed on the side of the frame, and a blade mounted on the single support, characterized in that, The cutting tool is driven by a multi-degree-of-freedom adjustment mechanism with two-dimensional linkage capability. The adjustment mechanism includes: The first linear motion unit is set along the feeding direction of the sheet metal; A second linear motion unit is positioned along the thickness direction of the sheet material; The device further includes: Sensor components used for real-time detection of the edge position, tilt angle, and rate of change of sheet metal; A trajectory planning and control system electrically connected to the first linear motion unit, the second linear motion unit, and the sensor assembly; wherein... The control system is configured to generate a continuous spatial target trajectory based on the edge detection data of the board material, including a front-end trimming trajectory, a stable segment trimming trajectory, and a rear-end trimming trajectory. The control system generates a transition curve between the front-end trimming and the back-end trimming through trajectory interpolation, enabling the two linear motion units to execute in a continuous time manner, thereby completing all the trimming tasks of a traditional double-blade system under a single-blade structure. The control system predicts the back-end switching time based on the plate length and feed speed, enabling the tool to perform continuous processing of front-end trimming, upper and lower edge trimming, and back-end trimming without changing the number of supports and the number of tools.

2. The adaptive single-blade trimming device according to claim 1, characterized in that, The first linear motion unit is driven by a servo linear motor, and the second linear motion unit is driven by a stepper motor or a servo motor. The two motion units are controlled by independent drive controllers to improve the dynamic response capability during trajectory interpolation.

3. The adaptive single-blade trimming device according to claim 1, characterized in that, The sensor assembly includes a distance sensor, an angle sensor, and a wheel position detector. The control system performs fusion calculations based on the data from the three sets of sensors to obtain the real-time offset, local curvature, and rate of change of the edge of the board.

4. The adaptive single-blade trimming device according to claim 1, characterized in that, The trajectory planning and control system includes a trajectory generation module, a trajectory interpolation module, and a speed prediction module. The trajectory generation module is used to generate a first continuous spatial trajectory based on the geometry of the front end of the sheet metal. The trajectory interpolation module is used to correct the tool trajectory based on the real-time offset during the stable trimming stage. The speed prediction module is used to adjust the acceleration constraints of the trajectory based on the changes in the sheet metal feed speed.

5. The adaptive single-blade trimming device according to claim 1, characterized in that, When the control system detects that the tail of the board is about to enter the trimming area, it calculates the optimal switching time for the rear trimming based on the board length, current speed, and acceleration change trend, and generates a second continuous spatial trajectory to complete the rear trimming.

6. The adaptive single-blade trimming device according to claim 1, characterized in that, The maximum acceleration ratio of the first linear motion unit and the second linear motion unit of the multi-degree-of-freedom adjustment mechanism satisfies A1 ≥ 2·A2, so as to ensure the response speed of the tool during the front and rear end switching stage.

7. The adaptive single-blade trimming device according to claim 1, characterized in that, When performing front-end trimming, the control system first drives the second linear motion unit to perform pre-contact depth compensation, and then the first linear motion unit enters the tracking state, thereby reducing front-end cutting vibration.

8. The adaptive single-blade trimming device according to claim 1, characterized in that, When performing rear-end trimming, the control system first reduces the feed depth of the second linear motion unit, and then executes the tail cutting trajectory of the first linear motion unit, so that the tool maintains a stable cutting state the moment the plate leaves the template.

9. The adaptive single-blade trimming device according to claim 1, characterized in that, The continuous spatial trajectory includes a three-dimensional curved surface projection trajectory covering the upper edge, lower edge, and rounded corner areas of the plate, enabling a single blade to complete the trimming of multiple parts within a single mechanism.

10. A single-blade tracking-based edge-sealing and trimming method, applied to the adaptive single-blade trimming device of claim 1, characterized in that, The method includes: (1) Collect the feed speed and initial position signals of the board when it enters the edge banding machine, and start the sensor assembly; (2) Calculate the real-time offset, local curvature and rate of change of the plate edge based on the laser ranging, angle detection and wheel feedback data; (3) The trajectory planning and control system generates a continuous spatial trajectory including the front end, straight section and the rear end; (4) Control the first linear motion unit to execute the trajectory along the plate direction, and control the second linear motion unit to perform cutting depth compensation; (5) In the front-end region, the trajectory is triggered to enter the straight section based on the rate of curvature change and depth pre-compensation; (6) Perform dynamic lateral compensation and speed matching based on real-time edge tracking in straight sections; (7) Before the predicted end of the plate arrives, execute the end trimming trajectory and smoothly retract the tool according to the feed rate and cumulative processing time.

Images