Corner Matching in a Multi-RobotSewing
Automated sewing is a complicated task in man-ufacturing. Due to the non-rigid work pieces and variationsin the material characteristics, sensor-based control has to beused to accomplish the sewing operation. This paper presentsa strategy for velocity synchronization and corner matching inan automated sewing cell based on two industrial manipulatorsand a sewing machine. A hybrid force/motion control schemeis adopted using feedback from force/torque sensors for tensioncontrol and optical sensors to control the seam position. Thestrategy is based on switching between force control and dis-placement control using a leader/follower coordination scheme.This addresses the problem of corner mismatch occurring whentwo independent force controllers are used for controlling thetwo robots. Experiments verify that the proposed method gives asatisfactory corner matching, which is crucial for the presentedsewing case.The automation of sewing operations is a dif cult taskin manufacturing. Several challenges arise from the non-rigidity and the large variations and uncertain material char-acteristics of the processed materials. This makes the designand implementation of both material handling and controlduring the sewing applications a complex task compared tohandling rigid materials. Nevertheless, there is a demandfrom the industry in high-cost countries to automate thesewing process.In the past decades, several research groups have workedin the eld of automated sewing and the handling of non-rigid materials.An automated sewing cell consisting of one robot and asewing machine is presented in [1], [2]. Both the tensionin the work piece and the seam allowance are controlled inreal-time.In [3], a device for handling curved fabrics during a sewingoperation is presented. It is based on rollers in front ofand behind the needle. Different feeding speeds allow foradapting to different seam lengths.In [4], an overview of the challenge of automated sewingis presented, especially the sewing of 3D-shaped products.The focus is on material handling. Adaptive control strategiesbased on measurements of the seam allowance and the feerate during the sewing operation are suggested.A system based on two robots is presented in [5]. Therobots work together to handle a single piece of fabric duringthe sewing operation. The system includes controllers forpressing force and tension in the fabric. The position iscontrolled by a visual tracking of the fabric.Another sewing cell demonstrator based on a single robotand a sewing machine is presented in [6]. The task is to sewan assembly of two similarily shaped parts. A triangulation-based sensor is used for edge detection while optical motionsensors are used for measurement of the sewing speed.A demonstrator based on a sewing machine with a servo-controlled feeding mechanism is presented in [7]. Two parts,that are separated by a thin plate, are controlled indepen-dently by the servo mechanism. An open loop path controlis used which utilizes recognition of patterns on the fabric.The need for sensor-based feedback control is emphasizedby the authors.VS Sewing Machines
The contribution of this paper is a new technique forcorner matching and experiments verifying the feasibility ofthe proposed method. Further, this paper includes a moredetailed presentation of the control mechanism, especiallywith focus on corner matching.Based on the experience from the past work, a new controlstrategy for sewing of two parts is proposed. This methodis based on stretching the material in order to align theend points of the edges. Technically, this method uses aleader/follower coordination strategy. The robot with thesmaller force vector in needle direction acts as leader andis in force control mode, while the robot with the largerforce is controlled to keep the same distance to the needleas the leader. The possible switching between leader andfollower is evaluated frequently, for example once a second.The force set point for the leader robot is set to2 N, whichprevents the parts from wrinkling. This strategy results ina sewing force larger than the ideal force set point, as atrade-off for sewing speed synchronization. This is differentfrom the theoretical approaches in [11], where the force isin uenced by pulses in the robot movement or by controllingthe feeding mechanism.The edge controllers are not in uenced by this techniqueand act independently for the two robots.To make this method work, several conditions have to bemet: The edge lengths of the two parts have to be similar.If the difference is too large, the parts cannot be sewnwhile keeping the sewing force within an appropriatelevel. The individual material characteristics of the two partshave to be equal within a certain range in order to obtainsimilar feed rates. Small differences are compensated bythe proposed control mechanism. The individual gripping points have to be in the pointsthat are to be matched, for example the corner of theparts. The seam segment has to be nearly straight so thegripping points are located nearly the extension of thesewing line. Curved parts can be sewn by partitioningthem into a sequence of nearly straight segments.When these conditions are not met, there is a risk of anincrease of the sewing force in one of the parts to a value thathas negative in uence on the sewing process. To prevent this,the sewing force of the distance-controlled robot is recorded,and in case a maximum force threshold is exceeded thesystem can be stopped for manual inspection of the errorcondition.
https://www.vssewingmachine.in/
Automated sewing is a complicated task in man-ufacturing. Due to the non-rigid work pieces and variationsin the material characteristics, sensor-based control has to beused to accomplish the sewing operation. This paper presentsa strategy for velocity synchronization and corner matching inan automated sewing cell based on two industrial manipulatorsand a sewing machine. A hybrid force/motion control schemeis adopted using feedback from force/torque sensors for tensioncontrol and optical sensors to control the seam position. Thestrategy is based on switching between force control and dis-placement control using a leader/follower coordination scheme.This addresses the problem of corner mismatch occurring whentwo independent force controllers are used for controlling thetwo robots. Experiments verify that the proposed method gives asatisfactory corner matching, which is crucial for the presentedsewing case.The automation of sewing operations is a dif cult taskin manufacturing. Several challenges arise from the non-rigidity and the large variations and uncertain material char-acteristics of the processed materials. This makes the designand implementation of both material handling and controlduring the sewing applications a complex task compared tohandling rigid materials. Nevertheless, there is a demandfrom the industry in high-cost countries to automate thesewing process.In the past decades, several research groups have workedin the eld of automated sewing and the handling of non-rigid materials.An automated sewing cell consisting of one robot and asewing machine is presented in [1], [2]. Both the tensionin the work piece and the seam allowance are controlled inreal-time.In [3], a device for handling curved fabrics during a sewingoperation is presented. It is based on rollers in front ofand behind the needle. Different feeding speeds allow foradapting to different seam lengths.In [4], an overview of the challenge of automated sewingis presented, especially the sewing of 3D-shaped products.The focus is on material handling. Adaptive control strategiesbased on measurements of the seam allowance and the feerate during the sewing operation are suggested.A system based on two robots is presented in [5]. Therobots work together to handle a single piece of fabric duringthe sewing operation. The system includes controllers forpressing force and tension in the fabric. The position iscontrolled by a visual tracking of the fabric.Another sewing cell demonstrator based on a single robotand a sewing machine is presented in [6]. The task is to sewan assembly of two similarily shaped parts. A triangulation-based sensor is used for edge detection while optical motionsensors are used for measurement of the sewing speed.A demonstrator based on a sewing machine with a servo-controlled feeding mechanism is presented in [7]. Two parts,that are separated by a thin plate, are controlled indepen-dently by the servo mechanism. An open loop path controlis used which utilizes recognition of patterns on the fabric.The need for sensor-based feedback control is emphasizedby the authors.VS Sewing Machines
The contribution of this paper is a new technique forcorner matching and experiments verifying the feasibility ofthe proposed method. Further, this paper includes a moredetailed presentation of the control mechanism, especiallywith focus on corner matching.Based on the experience from the past work, a new controlstrategy for sewing of two parts is proposed. This methodis based on stretching the material in order to align theend points of the edges. Technically, this method uses aleader/follower coordination strategy. The robot with thesmaller force vector in needle direction acts as leader andis in force control mode, while the robot with the largerforce is controlled to keep the same distance to the needleas the leader. The possible switching between leader andfollower is evaluated frequently, for example once a second.The force set point for the leader robot is set to2 N, whichprevents the parts from wrinkling. This strategy results ina sewing force larger than the ideal force set point, as atrade-off for sewing speed synchronization. This is differentfrom the theoretical approaches in [11], where the force isin uenced by pulses in the robot movement or by controllingthe feeding mechanism.The edge controllers are not in uenced by this techniqueand act independently for the two robots.To make this method work, several conditions have to bemet: The edge lengths of the two parts have to be similar.If the difference is too large, the parts cannot be sewnwhile keeping the sewing force within an appropriatelevel. The individual material characteristics of the two partshave to be equal within a certain range in order to obtainsimilar feed rates. Small differences are compensated bythe proposed control mechanism. The individual gripping points have to be in the pointsthat are to be matched, for example the corner of theparts. The seam segment has to be nearly straight so thegripping points are located nearly the extension of thesewing line. Curved parts can be sewn by partitioningthem into a sequence of nearly straight segments.When these conditions are not met, there is a risk of anincrease of the sewing force in one of the parts to a value thathas negative in uence on the sewing process. To prevent this,the sewing force of the distance-controlled robot is recorded,and in case a maximum force threshold is exceeded thesystem can be stopped for manual inspection of the errorcondition.
https://www.vssewingmachine.in/
