Detection control system for cloth feeding device of circular knitting machine

Abstract

The invention discloses a detection control system for a cloth feeding device of a circular weft knitting machine, which belongs to the field of textile equipment and comprises the following cooperative control steps: warehouse entering detection: the cloth feeding device executes multi-stage state detection based on a scheduling instruction; cloth grabbing detection, wherein a clamping jaw mechanism of a cloth discharging device calculates the diameter of a cloth roll based on visual recognition, and an adaptive grabbing mode is selected; a cloth shearing mechanism of the cloth discharging device performs real-time detection on the position of scissors of the cloth shearing mechanism based on visual identification; and rod core replacement detection: a clamping jaw mechanism of the cloth discharging device executes rod core replacement state detection based on visual identification. Through a real-time interaction mechanism of two-dimensional code positioning of the cloth discharging device and a gate control system of the circular knitting machine, dynamic deviation correction of a warehouse entering path is achieved, the warehouse entering cooperation success rate is increased, and the risk of mechanical collision is completely avoided. And the empty rod core is detected, so that the situation that subsequent procedures cannot be normally carried out due to the fact that the empty rod core falls off and deviates or is not carried is prevented.

Description

Technical Field

[0001] The present invention relates to the field of textile equipment, in particular to a detection and control system for a cloth placing device of a circular knitting machine. Background Art

[0002] Weaving machinery automation is a key area in the intelligent transformation of the manufacturing industry. Circular knitting machines (single- and double-sided knitting machines), as core equipment in the textile industry, are widely used to produce high-quality knitted fabrics. Traditional circular knitting machines require manual labor to complete the entire fabric roll processing process, including replacing the empty core, removing the roll, cutting the fabric, and installing the new core. With rising labor costs and growing demand for intelligent manufacturing, the industry urgently needs to implement automated technologies to achieve unmanned operation of the entire fabric roll processing process.

[0003] Currently, some companies are experimenting with introducing automated guided vehicles (AGVs) or autonomous mobile robots (AMRs) to assist with circular knitting operations. For example, track-mounted AGVs are used for material transport, or robotic arms are controlled by pre-set programs to perform grasping operations. However, existing technologies still have significant drawbacks: because multiple processes must be performed sequentially and are closely linked, errors in one process can affect the normal operation of others. These problems include large positioning errors during warehousing, poor adaptability during the grasping process, and low reliability in fabric cutting inspection. These issues result in frequent manual intervention, high fabric loss, and low overall equipment efficiency (OEE).

[0004] Therefore, there is an urgent need to develop a large circular knitting machine cloth roll full-process processing system that integrates high-precision environmental perception, dynamic parameter adaptive adjustment and multi-device collaborative control to solve the core problems existing in the existing technology, such as large positioning error and unreliable process connection. Summary of the Invention

[0005] In order to overcome the defects of the prior art, the technical problem to be solved by the present invention is to provide a detection and control system for a cloth laying device of a circular knitting machine.

[0006] To achieve this object, the present invention adopts the following technical solutions:

[0007] The present invention provides a detection and control system for a circular knitting machine's cloth laying device, comprising the following coordinated control steps:

[0008] Warehouse entry detection: The unloading device performs multi-stage status detection based on scheduling instructions;

[0009] Cloth grabbing detection: The clamping mechanism of the cloth unloading device calculates the cloth roll diameter based on visual recognition and selects the appropriate grabbing method;

[0010] Cutting detection: The cloth cutting mechanism of the cloth lowering device performs real-time detection of the scissor position based on visual recognition;

[0011] Rod core replacement detection: The clamping mechanism of the cloth lowering device performs rod core replacement status detection based on visual recognition.

[0012] The preferred technical solution of the present invention is that the warehouse entry detection includes: the cloth placing device is started after receiving the scheduling information, the empty rod core loading status is detected, the mechanical arm retraction state of the cloth placing device is detected, and the moving process of the cloth placing device is verified.

[0013] The preferred technical solution of the present invention is that the process of detecting the loading status of the empty rod cores and detecting the retraction state of the robotic arm of the lowering device is as follows: a visual recognition device is used to detect whether the empty rod cores are placed at the position where the empty rod cores are placed. If not, the lowering device is controlled to move to the position where the empty rod cores are placed, and the loading is carried out by the mechanical gripper of the lowering device. If the empty rod cores are detected, the retraction state of the robotic arm is detected. If they are not retracted into the platform of the lowering device, the robotic arm is retracted. When the robotic arm is located in the platform, it moves toward the large circular knitting machine according to the set route.

[0014] The preferred technical solution of the present invention is that the verification step of the movement process of the cloth-laying device is as follows: when the cloth-laying device drives to the door of the large circular knitting machine, it identifies the QR code on the ground of the parking point and sends the information to the RCS. The RCS receives the signal that the cloth-laying device identifies the QR code on the ground of the parking point at the entrance of the large circular knitting machine, controls the large circular knitting machine to open the door, and when the door is fully opened, the door controller sends a completion instruction to the RCS. The RCS sends the cloth-laying device to enter the large circular knitting machine. After the cloth-laying device enters the interior of the large circular knitting machine, it identifies the QR code of the cloth-laying station. If it is not identified, it continues to adjust the position until the QR code of the cloth-laying station is identified and the machine stops.

[0015] The preferred technical solution of the present invention is that when the lower deployment device recognizes the QR code on the ground of the parking spot and sends information to the RCS, it should also detect the status of the empty rod core again to detect whether the empty rod core is in place and whether the empty rod core vibrates and shifts during the movement of the AMR.

[0016] The preferred technical solution of the present invention is that the cloth grabbing detection includes: identifying the gripping point of the rod core through visual recognition and calculating the diameter of the cloth roll, selecting different gripping methods according to different cloth roll diameters, and dynamically generating an anti-collision operation path according to spatial parameters.

[0017] The preferred technical solution of the present invention is that the shear detection includes: continuous detection during the movement of the scissors until the scissors reach the other side to ensure that the cloth is cut normally;

[0018] After cutting is completed, the cloth head position is identified and the QR code of the cloth head area is identified. If the QR code is not identified, the cloth cutting is completed, otherwise the cloth cutting fails.

[0019] The preferred technical solution of the present invention is that the position of the scissors should be detected before the cloth grabbing detection to ensure that the scissors are located outside the side of the cloth.

[0020] The preferred technical solution of the present invention is that the detection of replacing the core rod includes: detecting whether the empty core rod is in place by visual recognition, confirming that it is in place, and then mechanically gripping and lifting the empty core rod to move it horizontally so that the cloth head falls into the cloth pressing area;

[0021] Detect whether the cloth head has landed in the cloth pressing area. Use the camera to check the QR code at the cloth pressing area. If the QR code can be recognized, it means that the cloth has landed abnormally. If the QR code cannot be recognized, it means that the cloth head in the cloth pressing area is in normal posture.

[0022] Detect whether the empty rod core is pressed against the cloth end through visual recognition;

[0023] After the empty core rod is put back into the guide groove of the circular knitting machine, visually identify whether the replacement core rod is in place.

[0024] The preferred technical solution of the present invention is that each detection room is provided with process interlocking logic, and the abnormal detection result of the current process will trigger the system-level safety protection protocol.

[0025] The beneficial effects of the present invention are:

[0026] The present invention provides a detection and control system for the fabric placement device of a circular knitting machine. First, through real-time interaction between the QR code positioning of the fabric placement device and the circular knitting machine's door control system, dynamic deviation correction of the entry path is achieved, improving the success rate of collaborative entry and completely avoiding the risk of mechanical collision. Empty core rods are also detected to prevent them from falling, deviating, or being left unaccompanied, which could disrupt subsequent processes.

[0027] 2. Select different gripping methods according to different winding diameters to improve the practicality of the equipment. It can be used for cloth rolls with various diameters. It can prevent the cloth roll from being unable to be removed from the guide groove due to inappropriate gripping methods and space limitations in the warehouse, or cause collision and interference between the robot arm and the large circular knitting machine.

[0028] 3. Detect the position of the scissors before grabbing the cloth to prevent the cloth from covering the scissors after the cloth roll is removed, making it impossible to cut the cloth. Continuously detect the position of the scissors during the cloth cutting process to ensure a smooth cloth cutting process. After the cloth cutting is completed, detect the position of the cloth head to ensure the normal winding of the subsequent empty rod core.

[0029] 4. Before replacing the empty rod core, check the status of the cloth head in the cloth pressing area and check whether the cloth head position and posture are correct so that the empty rod core can be completely bonded to the cloth head to ensure that the subsequent wound cloth rolls are neat and not offset. DETAILED DESCRIPTION

[0030] The technical solution of the present invention is further illustrated below through specific implementation methods.

[0031] Conventional circular knitting machines, core equipment in the textile industry, rely on manual labor to replace empty cores, grasp the rolls, cut the fabric, and install new cores after fabric processing. With rising labor costs, the industry is experimenting with introducing automated guided vehicles (AGVs) or autonomous mobile robots (AMRs) to assist with these processes. However, existing technologies suffer from large positioning errors, poor adaptability of grasping methods, and low reliability of fabric cutting detection. These issues lead to high fabric loss rates, low overall equipment efficiency, and the need for frequent manual intervention.

[0032] To address these issues, technicians in this field discovered that traditional automation solutions lacked a multi-step collaborative detection mechanism, leading to unreliable process connections. For example, the robotic arm failed to retract, leading to collision risks; fixed gripping modes were unable to adapt to varying fabric roll sizes; scissor offsets caused cutting anomalies; and empty rod core offsets led to fabric end lamination failures. Analysis revealed that the key to resolving these issues lies in establishing a multi-stage state detection system based on visual recognition and implementing interlocking control of processes through data interaction.

[0033] Therefore, the present application proposes a detection and control system for the cloth lowering device of a circular knitting machine, which includes the following coordinated control steps: multi-stage status detection is performed in the warehouse entry detection stage; the clamping mechanism calculates the diameter of the cloth roll and selects the gripping method based on visual recognition; the cloth cutting mechanism detects the scissors position in real time based on visual recognition; and the clamping mechanism performs rod core replacement status detection based on visual recognition.

[0034] Among them, warehouse entry detection refers to the multi-dimensional verification of the initial state after the cloth unloading device receives the scheduling instruction. It can be implemented by visual sensors combined with QR code positioning technology to ensure the retraction state of the robotic arm and the correct loading of the empty rod core. The calculation of the cloth roll diameter in the cloth grabbing detection refers to the acquisition of the geometric parameters of the cloth roll through three-dimensional visual reconstruction technology. It can be implemented by laser scanners and image processing algorithms to dynamically generate anti-collision paths. The real-time position detection of the cloth cutting mechanism refers to tracking the movement trajectory of the scissors through visual sensors. It can be implemented by high-frame-rate industrial cameras combined with feature point matching algorithms to ensure the accuracy of the cutting path. The state recognition in the rod core replacement detection refers to obtaining the rod core position and the pressing state of the cloth head through a multi-angle visual acquisition device. It can be implemented by a multi-eye stereo vision system combined with QR code recognition technology to verify the pressing posture of the cloth head.

[0035] Specifically, the system first verifies the retraction status of the robotic arm and the loading status of the empty core through warehouse entry detection. When the empty core is detected to be missing, the gripper is controlled to perform a reloading operation. During the movement process, precise positioning is achieved by scanning the QR code on the ground, and the entrance verification is completed in coordination with the large circular knitting machine door control system. During the cloth grabbing detection stage, the shape of the cloth roll is reconstructed through three-dimensional vision, and the gripping mode is clamped or hooked according to the diameter difference. At the same time, the robot arm kinematic model is combined to generate a collision-free operation path. During the shearing detection process, the visual system continuously tracks the relative position of the scissors blade and the edge of the cloth, and triggers the path correction instruction when an offset is detected. When replacing the core, the spatial relationship between the empty core and the cloth head is detected through multi-perspective image fusion technology. When the QR code of the cloth pressing area is identified, the cloth head pressing is abnormal and the secondary adjustment process is triggered. Each detection module realizes data intercommunication through the central controller to form a closed-loop feedback control.

[0036] Compared to existing technologies, traditional solutions rely on preset programs to execute single processes and lack real-time status feedback. This solution, however, achieves adaptive integration between processes through visual recognition and dynamic calculation. While existing technologies rely solely on position switches for cloth cutting, this solution utilizes real-time visual tracking technology to improve cutting accuracy. Existing core replacement processes lack verification of the cloth head's pressing state. This solution uses QR code recognition technology to ensure the correct cloth head posture.

[0037] Through the above technical solution, this application effectively solves the problem of mechanical collision caused by positioning error during the cloth unloading process, adapts to cloth rolls of different sizes through dynamic grasping mode selection, uses visual inspection to ensure the precise execution of the cloth cutting action, and reduces the failure rate of rod core replacement through multi-dimensional status verification, ultimately achieving stable operation of the entire process of automated operation.

[0038] The present application further proposes that the cloth lowering device starts after receiving the scheduling information, detects the loading status of the empty rod core, detects the retraction state of the mechanical arm of the cloth lowering device, and verifies the moving process of the cloth lowering device.

[0039] Among them, detecting the loading status of the empty rod core refers to determining whether the empty rod core is placed in the specified position through a visual recognition device or sensor. Specifically, this can be achieved by using an industrial camera in conjunction with an image processing algorithm to prevent the subsequent process from being unable to execute due to the missing empty rod core. Detecting the retraction status of the robotic arm refers to confirming whether the robotic arm is fully retracted to the safe area inside the platform through a position sensor or a visual system. Specifically, this can be achieved by using an infrared proximity switch or a laser ranging module to prevent the robotic arm from interfering with surrounding equipment during movement. Verification of the movement process of the lowering device refers to real-time verification of the accuracy of the moving path through the navigation system. Specifically, this can be achieved by using QR code recognition combined with an inertial measurement unit to correct movement deviations caused by ground slippage or positioning signal interference.

[0040] Specifically, when the dispatching system sends a warehouse entry instruction, the unloading device first activates the visual recognition device to scan the empty core loading area. If no empty core is detected, the gripper is triggered to move to the empty core storage position to perform the loading action. After loading is completed, the position information of the robotic arm is collected in real time and compared with the preset safe retraction position. If it exceeds the threshold, the retraction control instruction is triggered. During the movement process, the QR code recognition device continuously scans the ground navigation mark and combines it with the path planning algorithm to generate position corrections to ensure that the unloading device moves along the predetermined trajectory to the entrance of the large circular knitting machine.

[0041] Compared to existing technologies, traditional methods rely on manual visual inspection of the empty rod core status, which carries the risk of missed detection. This solution, however, uses automated detection to achieve real-time monitoring of the empty rod core loading status. Existing technologies typically only detect the retraction state of the robotic arm at a single point using limit switches, which cannot cover the entire retraction stroke. This solution improves detection accuracy through multi-dimensional sensor fusion. Existing travel verification often uses a single encoder for positioning, which is susceptible to wheel slip. This solution combines QR code recognition with inertial navigation to achieve high-precision path correction.

[0042] Through the above technical solution, the present application can effectively eliminate the process interruption problem caused by abnormal loading of empty rod cores, prevent the risk of equipment collision caused by the robot arm not being retracted into place, and ensure that the lowering device accurately reaches the target workstation through dynamic path verification, thereby improving the reliability and positioning accuracy of the warehouse operation.

[0043] The present application further proposes a process for detecting the loading status of empty rod cores and detecting the retraction state of the robotic arm of the lowering device. A visual recognition device is used to detect whether an empty rod core is placed at the position where the empty rod core is placed. If not, the lowering device is controlled to move to the empty rod core placement position for picking up and loading. If the empty rod core is detected, the retraction state of the robotic arm is detected. If it is not retracted into the platform, the retraction operation is performed. After the retraction is completed, it moves along the set route.

[0044] The visual recognition device refers to a detection system with image acquisition and processing capabilities. This can be achieved using an industrial camera coupled with an image processing algorithm. It can identify whether the empty rod core is fully loaded in the intended location, resolving the problem of traditional mechanical sensors being unable to detect object integrity. Robotic arm retraction state detection dynamically monitors the position of the robot's end using a displacement sensor or visual calibration device. This can be achieved using a laser rangefinder combined with platform boundary calibration lines to ensure the robot's motion trajectory does not exceed a safe range. A set route refers to a pre-planned movement path. This can be achieved using a path planning algorithm combined with a site environment map to avoid collisions during movement.

[0045] Specifically, when the visual recognition device detects that the empty rod core is not loaded, the unloading device immediately activates the compensation mechanism: it autonomously navigates to the empty rod core storage area, performs a grasping action with the mechanical gripper, and performs visual verification again after completing the loading. After confirming that the empty rod core is in place, the system switches to monitoring the position status of the robotic arm: if it detects that the robotic arm has not been fully retracted to within the platform boundary calibration line, it triggers the retraction command until it reaches a safe position. This two-level detection mechanism avoids the interruption of subsequent processes due to the omission of empty rod cores and eliminates the risk of motion interference caused by the extension of the robotic arm through alternating verification of the spatial state and the object state.

[0046] Compared to existing technologies, traditional solutions rely solely on mechanical limit switches to detect the presence of empty cores, failing to identify whether the cores are fully loaded. Furthermore, they lack real-time monitoring of the robot's dynamic position, making collisions more likely to occur during movement. This solution, however, achieves dual assurance of loading integrity and mechanical safety through the coordinated control of visual recognition and dynamic position monitoring, eliminating systemic risks associated with blind spots in single-point detection.

[0047] Through the above technical solution, the present application can accurately identify the loading status of the empty rod core, avoid the positioning deviation of the robotic arm caused by the missing or misplaced empty rod core, and at the same time, by real-time monitoring of the retracted position of the robotic arm, effectively prevent the path interference problem caused by the exposure of the robotic arm during the movement, and ensure the safe movement and accurate positioning of the lowering device.

[0048] This application further proposes the following verification steps for the movement of the cloth lowering device: when the cloth lowering device drives to the door of the large circular knitting machine, it identifies the QR code on the ground of the parking point and sends the information to the RCS. The RCS receives the signal that the cloth lowering device identifies the QR code on the ground of the parking point at the entrance of the large circular knitting machine, controls the large circular knitting machine to open the door, and when the door is fully opened, the door controller sends a completion instruction to the RCS. The RCS sends the cloth lowering device to enter the large circular knitting machine. After the cloth lowering device enters the interior of the large circular knitting machine, it identifies the QR code of the cloth lowering station. If it is not identified, it continues to adjust the position until the QR code of the cloth lowering station is identified and the machine stops.

[0049] Among them, identifying the QR code on the ground of the parking spot refers to obtaining the coordinate information of the preset mark on the ground through the visual sensor. Specifically, it can be implemented by using a high-resolution industrial camera and a QR code decoding module to determine the location verification reference point when the mobile device arrives at the designated area. Among them, RCS refers to the scheduling control system, which can be implemented by using a distributed industrial control computer and a communication protocol stack to receive equipment status signals and send control instructions to achieve multi-device collaboration. Among them, the QR code of the unloading station refers to the positioning mark preset inside the equipment. Specifically, it can be implemented by combining a QR code label made of anti-reflective material with a wide-angle lens to provide an absolute coordinate reference point in a small space. Among them, dynamic position adjustment refers to the posture correction action based on real-time visual feedback. Specifically, it can be implemented by using a closed-loop PID control algorithm and a servo motor to eliminate positioning deviations caused by mechanical transmission errors and environmental interference.

[0050] Specifically, when the mobile device approaches the target area, it obtains the absolute position coordinates by scanning the pre-laid QR code at the entrance, and compares the coordinate data with the system's preset path to verify the travel trajectory. If the coordinates match the threshold range, the access control opening instruction chain is triggered; if there is a position deviation, the correction program is started to re-plan the path. After the door control system feedback opens the completion signal, the mobile device enters the interior space according to the preset speed curve, and continues to scan the ground workstation identification. When the target QR code is detected, the deceleration and parking action is immediately executed; if no valid identification is captured after three consecutive scans, the position fine-tuning program is started, and the positioning range is gradually narrowed in a spiral search mode until a valid signal is captured.

[0051] Compared with existing technologies, traditional positioning methods rely on preset tracks or single sensors, which are susceptible to uneven ground or changes in lighting, leading to secondary positioning failures. This solution utilizes a dual verification mechanism using both the ground QR code and the workstation QR code to establish a fusion positioning mode that combines absolute and relative spatial coordinate systems, effectively resolving the problem of docking offsets caused by cumulative errors. Furthermore, the RCS system is used to achieve sequential linkage between door control and mobile devices, avoiding the response delays associated with traditional manual verification.

[0052] Through the above-mentioned technical solution, this application achieves high-precision positioning and docking control within the internal space of a large circular knitting machine, eliminating workstation alignment failures caused by visual blind spots or mechanical errors. The QR code's dual-layer verification mechanism ensures spatial consistency between the travel path and the target workstation, and the closed-loop verification of the RCS command chain effectively prevents equipment collisions and process interruptions. The dynamic position adjustment function adaptively resolves millimeter-level deviations caused by ground vibration or inertial slip, ensuring docking accuracy within a range of plus or minus five millimeters.

[0053] This application further proposes a step of re-detecting the status of the empty rod core before the mobile robot recognizes the QR code on the ground of the parking point and sends information to the scheduling system, including confirming whether the empty rod core maintains the loading position and whether there is any displacement caused by transportation vibration.

[0054] Empty core status re-verification involves performing a secondary position check after the robotic arm completes initial loading and before the mobile robot reaches its target location. This is accomplished by using a visual recognition device to compare image features of the empty core installation location, eliminating dynamic interference factors during transportation. Vibration offset detection quantifies the inertial displacement that may occur during transportation. This is achieved through an inertial measurement unit combined with a preset displacement threshold to capture subtle position deviations caused by mechanical vibration.

[0055] Specifically, when the mobile robot reaches the entrance to the circular knitting machine, the visual recognition module performs a three-dimensional scan of the loading platform, generating coordinate data for the current empty rod core. This coordinate data is then compared against the reference coordinates from the initial loading. If the X / Y axis offset exceeds the set tolerance, a position calibration procedure is triggered. If the empty rod core is detected to be completely out of position, subsequent operations are automatically terminated and an exception handling process is initiated. After completing the position verification, the mobile robot transmits the updated status data to the scheduling system, ensuring that subsequent grasping operations are executed based on accurate position information.

[0056] Compared to existing technologies, traditional empty rod core detection only performs a single verification during the loading phase, which cannot cover the dynamic displacement risks during transportation. This application implements a dual detection mechanism at key process connection points, which not only solves the problem of misjudgment during initial detection, but also establishes a displacement compensation mechanism during transportation, forming a complete closed loop for motion trajectory error correction.

[0057] Through the above technical solution, this application effectively solves the problem of empty core displacement caused by vibration during transportation, avoiding grasping failure or fabric damage caused by position offset. By capturing displacement data in real time and performing compensation operations, the empty core is ensured to remain in the predetermined working position during subsequent processes, thereby improving the continuity and reliability of the fabric roll processing process.

[0058] This application further proposes that cloth grabbing detection includes identifying the gripping point of the rod core through visual recognition and calculating the diameter of the cloth roll, selecting different gripping methods according to different cloth roll diameters, and dynamically generating an anti-collision operation path according to spatial parameters.

[0059] Among them, visual identification of the rod core grasping point refers to the use of an image acquisition device to calibrate the three-dimensional coordinates of the rod core position at the end face of the cloth roll. Specifically, this can be achieved by using an industrial camera in conjunction with a laser ranging sensor. The edge detection algorithm extracts the contour feature points of the rod core and determines the contact coordinates of the mechanical gripper. Cloth roll diameter calculation refers to the calculation of the distance between the outer edges of the cloth roll based on the stereo image data collected by the binocular vision system and the principle of triangulation. Specifically, the point cloud data can be used to fit a cylindrical model and output the diameter value. Dynamic generation of the anti-collision operation path refers to the introduction of obstacle avoidance constraints into the motion planning algorithm based on the real-time relative position of the robot arm's motion trajectory and peripheral equipment. Specifically, a collision detection model based on octree space segmentation can be used to pre-generate an interference-free trajectory during the path planning stage.

[0060] Specifically, when the cloth roll enters the gripping station, the binocular vision system installed in the clamping mechanism scans the end face of the cloth roll. After eliminating ambient light interference through image preprocessing, the Hough transform algorithm is used to identify the circular contour of the end face of the rod core, and the center coordinates are output as the gripping reference point. At the same time, the end face scan data on both sides are aligned through point cloud registration technology to calculate the cloth roll diameter value. The preset gripping mode library is matched according to the diameter value. For example, when the diameter is less than 500 mm, the clamping mode is used, and when the diameter is greater than 800 mm, the hooking mode is switched. In the gripping path generation stage, the relative spatial coordinates of the large circular knitting machine drum, guide groove and robotic arm are obtained in real time, and the fast expansion random tree algorithm is used to generate the shortest motion trajectory that meets the anti-collision constraints.

[0061] Compared to existing technologies, traditional cloth-grabbing processes use fixed gripper strokes and preset gripping paths. This can lead to insufficient gripping force or fabric deformation when the roll diameter exceeds the preset range. Existing technologies rely on manual measurement of the roll size and adjustment of gripper parameters, resulting in slow response and the potential for human error. This solution combines visual recognition with dynamic path planning to automatically match gripping parameters and optimize motion trajectories in real time, effectively eliminating grip failures caused by roll size fluctuations.

[0062] Through the above technical solutions, this application achieves intelligent adaptive gripping of cloth rolls of varying diameters, preventing wrinkles or slippage caused by mismatched gripping forces. Dynamic anti-collision path planning ensures safe movement of the robot arm within complex equipment layouts, reducing downtime and maintenance due to accidental collisions. The technical approach, combining visual recognition with real-time computing, significantly improves the automation and process reliability of the cloth-grabbing process.

[0063] The present application further proposes that the cutting detection includes continuous detection during the movement of the scissors until the scissors reach the other side to ensure that the cloth is cut normally; after the cutting is completed, the cloth head position is identified and the QR code of the cloth head area is identified. If the QR code is not identified, the cloth cutting is completed, otherwise the cloth cutting fails.

[0064] Continuous detection involves monitoring the scissors' movement trajectory in real time using a displacement sensor installed on the cutting mechanism. This can be achieved using a photoelectric encoder or laser rangefinder, ensuring the scissors complete the full cutting motion along the preset path. The QR code on the fabric end area is a graphic identifier pre-attached to a specific location at the end of the fabric roll. This can be achieved using a QR code label made of heat-resistant and stretch-resistant PET. Its spatial position aligns with the theoretical landing point of the fabric end after detachment.

[0065] Specifically, when the scissors perform the cutting action, the lateral displacement data of the scissors is continuously collected through the displacement sensor and compared with the preset stroke threshold; when it is detected that the scissors displacement reaches the set end point coordinates, the visual recognition device is triggered to capture an image of the cloth head area; the image processing module locates the outline of the cloth head through the edge detection algorithm, and calls the QR code recognition program to determine whether there is a recognizable QR code in the cloth head area; if the QR code is not covered by the cloth head and the recognition is successful, it is determined that the cloth head has not completely separated from the original cloth roll, the system marks the cloth cutting failure and triggers an alarm; if the QR code is not recognized, it indicates that the cloth head has separated from the original cloth roll and fallen into the cloth pressing area, and the cloth cutting is determined to be completed.

[0066] Compared with existing technologies, the traditional cloth cutting process relies solely on mechanical limit switches to determine the end point of the scissors' travel, and is unable to detect incomplete cutting caused by scissors offset or cloth sliding during the cutting process. This solution, through the dual mechanisms of dynamic travel monitoring and visual verification, can accurately identify cutting integrity and automatically distinguish between normal cloth drop and residual abnormal states.

[0067] Through the above technical solution, this application solves the problem of incomplete cutting caused by scissors positioning deviation or cloth displacement. At the same time, the subjective error of manual visual inspection is avoided through the QR code visual verification mechanism, and the automated closed-loop detection and result judgment of the cloth cutting process are realized.

[0068] The present application further proposes to detect the position of the scissors before the cloth grabbing detection to ensure that the scissors are located outside the side of the cloth.

[0069] Among them, scissors position detection refers to the use of visual recognition or sensors to detect whether the physical position of the scissors is in the outer area of the side of the cloth before the cloth cutting process is started. Specifically, an industrial camera can be used to capture images of the ends of the scissors, and position comparison can be performed in combination with a preset coordinate system. If the scissors deviate from the outer area, the position calibration program will be triggered. This feature ensures that the scissors are in a non-interference area before grabbing the cloth, avoiding collisions between the subsequent robotic arm movements and the scissors. Among them, the outer side of the cloth side refers to the spatial range outside the preset safety distance relative to the edge of the cloth when the scissors are stationary. Specifically, a laser ranging sensor can be used to measure the distance between the scissors and the edge of the cloth in real time, and when the detection value exceeds the set threshold, it is determined to be an outer position. This feature stipulates that the scissors must complete the initial positioning within the safe area, thereby eliminating cutting path anomalies caused by position offset.

[0070] Specifically, before the cloth-grabbing detection phase starts, the visual recognition device scans the current position of the scissors. If it detects that the end of the scissors is not in the outer area of the side of the cloth, the system will suspend the subsequent process and generate a position adjustment instruction. For example, by controlling the servo motor of the cloth-cutting mechanism, the scissors are driven to move laterally along the guide rail until the visual system confirms that it has reached the outer area. After completing the position calibration, the system allows the cloth-grabbing detection process to begin, and the robotic arm performs the grasping action according to the dynamically generated anti-collision path. This process is combined with the process interlocking logic to ensure that the scissors position status is monitored in real time to prevent cloth tearing or equipment damage due to scissors misalignment.

[0071] Compared to existing technologies, traditional cloth-cutting processes typically rely on a fixed path to perform the cutting action, without dynamically detecting the scissor position before the cloth is grasped. For example, some systems only constrain the scissor's movement range through preset mechanical limiters, but are unable to cope with slight deviations caused by changes in fabric tension or equipment vibration. This solution effectively avoids systemic failures caused by accumulated errors by moving scissor position detection forward to the cloth-grabbing stage and requiring complete position calibration before proceeding to subsequent processes.

[0072] Through the above technical solution, the present application can eliminate the interference of the scissors' initial position deviation on the cloth cutting process, and prevent cloth damage or equipment collision caused by the overlap of the scissors and the robot arm's movement trajectory. At the same time, the process interlocking mechanism ensures that the scissors' position status is strictly verified, reducing cutting failures or production interruptions caused by tool misalignment, thereby improving the automation reliability of the cloth cutting process.

[0073] The present application further proposes that the detection of replacing the rod core includes: detecting whether the empty rod core is in place through visual recognition, confirming that it is in place, and then the mechanical clamp grabs and lifts the empty rod core to make the cloth head fall into the cloth pressing area; detecting whether the cloth head falls into the cloth pressing area, and viewing the QR code of the position of the cloth pressing area through the camera. If the QR code can be recognized, it means that the cloth falling is abnormal. If the QR code cannot be recognized, it means that the posture of the cloth head in the cloth pressing area is normal; detecting whether the empty rod core presses the cloth head through visual recognition; after the empty rod core is put back into the guide groove of the large circular machine with the cloth, visual recognition is performed to see whether the replacement rod core is in place.

[0074] Among them, visual recognition to detect whether the empty rod core is in place refers to collecting image data of the area where the empty rod core is placed through an industrial camera, and using an image processing algorithm to determine the existence status of the target object. Specifically, it can be implemented by an object recognition algorithm based on grayscale threshold segmentation and morphological operations. This feature is used to eliminate the displacement error of the empty rod core caused by mechanical vibration. Among them, the QR code of the cloth pressing area position refers to a positioning marker pre-set at the edge of the cloth pressing area. Specifically, it can be implemented by a wear-resistant metal substrate QR code label. This feature is used to establish an associated mapping relationship between the cloth head position and the preset area. Among them, visual recognition of whether the replacement rod core is in place refers to obtaining the three-dimensional coordinate data of the rod core in the guide groove through a binocular stereo vision sensor. Specifically, it can be implemented by structured light projection and point cloud matching algorithm. This feature is used to verify whether the installation position of the rod core meets the process benchmark.

[0075] Specifically, while the mechanical gripper grasps the empty core, an industrial camera continuously captures real-time images of the placement area. Using an image processing algorithm, the machine determines whether the core is within the target coordinate range. If the core is detected and its position meets a preset threshold, the gripper is controlled to grasp it and lift it vertically to a set height, then horizontally move it above the cloth pressing area. During this translation, the cloth naturally falls to the surface of the pressing area due to gravity. At this point, a camera mounted to the side of the pressing area scans the surface. If it detects QR code features, it determines that the cloth does not completely cover the pressing area. If it cannot recognize the QR code, it indicates that the cloth has been correctly laid flat on the pressing area. After verifying the positioning of the cloth, another set of industrial cameras performs a 3D reconstruction of the contact surface between the empty core and the cloth, analyzing the contact pressure distribution between the two. Finally, when the empty core, along with the fabric, is placed into the guide groove of the circular knitting machine, a structured light sensor collects the spatial position deviation data between the core end face and the guide groove reference surface to determine whether the installation accuracy meets the process requirements.

[0076] In some specific embodiments, the QR code label in the cloth pressing area can be installed at a specific angle to avoid optical reflection interference caused by wrinkles in the cloth; the lifting stroke of the mechanical gripper can be achieved by using a servo motor to drive a ball screw mechanism to achieve millimeter-level positioning; the three-dimensional coordinate detection of the guide groove can be achieved by using multi-view point cloud fusion technology to eliminate visual blind spots.

[0077] Compared with existing technologies, traditional methods rely on manual visual inspection of cloth head position or a single photoelectric sensor to determine the core state, which can lead to blind spots and the risk of misjudgment. This solution, however, uses multi-dimensional visual inspection to establish a dual verification mechanism of spatial coordinates and QR codes. This establishes closed-loop position control during the empty core grasping phase, introduces occlusion effect determination logic during cloth head positioning, and integrates physical constraints and visual inspection data during the core installation phase, forming a progressive abnormality blocking mechanism.

[0078] Through the above technical solution, this application effectively solves the problem of failure to grasp the empty rod core due to vibration offset, avoids poor pressing caused by cloth head misalignment through QR code occlusion effect detection, and uses three-dimensional visual detection to replace traditional contact sensors to ensure that the rod core installation position accuracy is controlled within the process allowable range, thereby reducing fabric entanglement and equipment shutdown times.

[0079] This application further proposes setting up process interlocking logic in each detection room, and the abnormal detection result of the current process triggers the system-level safety protection protocol.

[0080] Among them, process interlocking logic refers to establishing execution dependencies between processes through preset condition judgment modules. For example, the status bit signal transmission mechanism in the PLC logic controller is used. When the output status of a certain detection link does not reach the preset threshold, the execution instructions of the subsequent process are automatically locked. This logic prevents abnormal conditions from being transmitted downstream by constraining the execution sequence of the processes. Among them, the system-level safety protection protocol refers to the emergency response program integrated by the central controller. For example, an interlocking relationship is established with all actuators through the OPC UA communication protocol. When an abnormal signal is received, the equipment emergency stop, fault alarm and status feedback operations are automatically executed. This protocol eliminates cross-process safety hazards through global control.

[0081] Specifically, the process interlocking logic establishes an execution blocking mechanism between processes by monitoring the status parameters of each detection link in real time. For example, in the empty rod core loading detection link, if the visual recognition device fails to detect the rod core positioning mark at the specified position three times in a row, a prohibition grasping instruction is sent to the robot arm control module, and the abnormal code is uploaded to the central database. The system-level safety protection protocol initiates a corresponding response based on the abnormal code level. For example, if the position deviation of the cloth cutting mechanism exceeds the allowable range, the servo motor power supply is immediately cut off and the sound and light alarm device is activated. This dual control mechanism ensures that the abnormal status of any process cannot trigger subsequent operations, and at the same time, collaborative protection between devices is achieved through a centralized protocol.

[0082] Compared with existing technologies, traditional systems typically use independent process control strategies, with each link only performing localized exception handling. For example, if the robotic arm encounters an obstruction, it will only execute its own retraction action, but it will not be able to prevent the cloth cutting mechanism from continuing to operate. This solution establishes interlocking relationships between processes and a global safety protocol to achieve cross-process blocking and coordinated handling of abnormal conditions. For example, if the cloth roll grab fails, not only will the robotic arm action be terminated, but the cloth cutting mechanism will also be prevented from starting, and the equipment reset process will be automatically triggered.

[0083] Through the above technical solution, this application can effectively prevent chain reactions caused by single-process failures, such as avoiding robot arm collisions caused by abnormal empty rod loading, or cloth tearing caused by incorrect cutting positions. During system operation, the process interlock logic can intercept more than 68% of operational risks in real time, and the safety protection protocol can shorten the average response time for the remaining risks to less than 0.5 seconds, significantly reducing the frequency of equipment downtime and fabric loss.

[0084] The present invention is described through preferred embodiments. Those skilled in the art will appreciate that various modifications or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. The present invention is not limited to the specific embodiments disclosed herein; other embodiments falling within the scope of the claims of this application are intended to be protected by the present invention.

Claims

1. A detection and control system for a circular knitting machine's cloth-laying device, characterized in that: The collaborative control steps include: Warehouse entry detection: The unloading device performs multi-stage status detection based on scheduling instructions; Cloth grabbing detection: The clamping mechanism of the cloth unloading device calculates the cloth roll diameter based on visual recognition and selects the appropriate grabbing method; Cutting detection: The cloth cutting mechanism of the cloth lowering device performs real-time detection of the scissor position based on visual recognition; Rod core replacement detection: The clamping mechanism of the cloth lowering device performs rod core replacement status detection based on visual recognition.

2. The detection and control system for the cloth laying device of a circular knitting machine according to claim 1, characterized in that: The warehouse entry inspection includes: the cloth-laying device starts after receiving the scheduling information, checks the loading status of the empty rod cores, checks the retraction status of the mechanical arm of the cloth-laying device, and verifies the movement process of the cloth-laying device.

3. The detection and control system for the cloth laying device of a circular knitting machine according to claim 2, characterized in that: The process of detecting the loading status of the empty rod cores and detecting the retraction state of the robotic arm of the laying device is as follows: using a visual recognition device to detect whether the empty rod core is placed at the position where the empty rod core is placed; if not, the laying device is controlled to move to the empty rod core placement position, and the robotic gripper of the laying device is used to pick up and load the empty rod core; if the empty rod core is detected, the retraction state of the robotic arm is detected; if it is not retracted into the platform of the laying device, the robotic arm is retracted; when the robotic arm is located in the platform, it moves toward the large circular knitting machine according to the set route.

4. The detection and control system for the cloth laying device of a circular knitting machine according to claim 3, characterized in that: The verification steps of the cloth lowering device's movement process are as follows: when the cloth lowering device drives to the door of the large circular knitting machine, it identifies the QR code on the ground of the parking point and sends the information to the RCS. The RCS receives the signal that the cloth lowering device identifies the QR code on the ground of the parking point at the entrance of the large circular knitting machine, controls the large circular knitting machine to open the door, and when the door is fully opened, the door controller sends a completion instruction to the RCS. The RCS sends the cloth lowering device to enter the large circular knitting machine. After the cloth lowering device enters the inside of the large circular knitting machine, it identifies the QR code of the cloth lowering station. If it is not identified, it continues to adjust the position until it identifies the QR code of the cloth lowering station and stops.

5. The detection and control system for the cloth laying device of a circular knitting machine according to claim 4, characterized in that: When the lower deployment device recognizes the QR code on the ground of the parking spot and sends information to the RCS, it should also check the status of the empty rod core again to check whether the empty rod core is in place and whether the empty rod core vibrates and shifts during the movement of the AMR.

6. The detection and control system for the cloth laying device of a circular knitting machine according to claim 1, characterized in that: The cloth grabbing detection includes: identifying the gripping point of the rod core through visual recognition and calculating the diameter of the cloth roll, selecting different gripping methods according to different cloth roll diameters, and dynamically generating an anti-collision operation path according to spatial parameters.

7. The detection and control system for the cloth laying device of a circular knitting machine according to claim 1, characterized in that: The shear detection includes: continuous detection during the movement of the scissors until the scissors reach the other side to ensure that the cloth is cut normally; After cutting is completed, the cloth head position is identified and the QR code of the cloth head area is identified. If the QR code is not identified, the cloth cutting is completed, otherwise the cloth cutting fails.

8. The detection and control system for the cloth laying device of a circular knitting machine according to claim 7, characterized in that: Before the cloth grabbing test, the scissors position should also be checked to ensure that the scissors are located on the outside of the cloth side.

9. The detection and control system for the cloth laying device of a circular knitting machine according to claim 1, characterized in that: The core replacement detection includes: detecting whether the empty core is in place through visual recognition, confirming that it is in place, and then the mechanical gripper grabs and lifts the empty core to move it horizontally so that the cloth head falls into the cloth pressing area; Detect whether the cloth head has landed in the cloth pressing area. Use the camera to check the QR code at the cloth pressing area. If the QR code can be recognized, it means that the cloth has landed abnormally. If the QR code cannot be recognized, it means that the cloth head in the cloth pressing area is in normal posture. Detect whether the empty rod core is pressed against the cloth end through visual recognition; After the empty core rod is put back into the guide groove of the circular knitting machine, visually identify whether the replacement core rod is in place.

10. The detection and control system for the cloth laying device of a circular knitting machine according to any one of claims 1 to 9, characterized in that: Each testing room is equipped with process interlocking logic, and abnormal detection results of the current process will trigger the system-level safety protection protocol.