Advanced Technologies for Discrete-Parts Manufacturing

A completely automatic, cost-effective, production of garments and other sewn items can be achieved. A novel approach that depends heavily on computing, high performance servo systems, and machine vision has been demonstrated. This proposal will lead to a the most critical element of such automation, a completely automatic sewing machine that will sew arbitrary but controlled paths in the two pieces of fabric being sewn together. Relatively modest extensions would lead to a work-cell consisting of several of these units, nearly contiguous, to make a simple garment such as a pair of jeans or a shirt completely automatically including the cutting of fabric. It is the flexibility of fabric that makes traditional sewing so difficult to automate. In the past, researchers have attempted to automate sewing based upon replicating human functions and building machines that sew based on Euclidean units of measure. A human cannot count threads as they sew, but today a robotic machine can be based on the very advanced state of sensors, particularly machine vision, and computation, to use a non-Euclidean measure, namely thread count. It is this innovative concept that will be exploited to automate the traditional sewing process.