The spark and crackle typically associated with new consumer electronics were subdued during 1975. Though the spotlight played briefly on video disks, little hardware is yet available — excitement dissipated quickly, awaiting the “big push” still scheduled for late this year when MCA/Philips and RCA start marketing players in the United States. Calculators were personal, portabl, and plentiful this past year, with new entries by Hewlett-Packard and Texas Instruments leading the pack. And jewelers, accustomed to the familiar tick of mechanical timepieces, had to adjust to the silence of the digital watch-which, as prices tumble, is showing up in chain store windows as well as posh display cases. Here, the technology tug-of-war revolves around readouts: light-emitting diodes (LEDs) vs. liquid crystal displays (LCDs).
To reduce the error in tension control, fuzzy logicand neuro-controllers were applied to a similar system by Panagiotis Koustoumpardis and Nikos Aspragathos . The objective was to meet the criteria set forth by Gershon with greater robustness and without quantitative understanding of fabric properties by using qualitative knowledge of fabric properties such as extensibility and flexibility to provide inputs to a neuro-controller. While the results showed improved robustness, the system involved a single piece of fabric and did not address the issue of attaching two pieces of fabric together.
No fabric control strategy to date has made a successful jump from prototype to a useful real-world method. Obstacles include the lack of an accurate and robust position control method, the difficulties arising from using a continuous robotic manipulator to maintain proper tension in the fabric as it is being fed by an incrementally applied actuator, and the focus of research on controlling a single sheet of fabric despite the necessity of controlling two sheets of fabric during the sewing process. In maintaining proper fabric tension, tension measurement is used to detect error in the feed rate of the robot relative to the dogs. The objective of the controller is to prevent the fabric from buckling due to compression or tension, which results in a poor quality seam. Tension measurement has not proven a method that is robust enough given the wide range of fabric properties.
The use of an overhead vision system and industrial manipulator, coupled with a standard sewing machine has not shown the ability to perform adequately regardless of the control strategy or trajectory generation used. The primary problem is a lack of robustness in the design based on deficiencies in the sensing methods and inaccuracy in the manipulators. In addition, no previous methods were developed to sew multiple sheets of fabric. The use of more precise sensors and actuators and more advanced control methods has been shown to increase the robustness but has not been entirely successful. New strategies are needed to meet the challenges of incorporating automation into garment manufacturing. This paper lays out a new methodology by which automated sewing may be realized. It focuses on a subset of automated sewing, namely the precise actuation and sensing of fabric near the sewing head.
Comments