Neural Adaptive Control for Positioning Fabric on a Frictional Surface

Abstract

This research focuses on real-time position control for draping fabric sliding on a high friction surface. Although fabrics are usually positioned on smooth surfaces with fixed fabric guides to simplify automated handling, a high friction work surface holds the fabric in place after positioning, allowing accurate assembly of multiple fabric parts without specialized jigs or fixtures. A neural adaptive controller with feedforward friction compensation provides asymptotic tracking for a spring mass model with friction. A test stand and an optical sensor are designed to facilitate real time position measurement and control. The neural adaptive controller demonstrates good position tracking and robustness to fabric property variations relative to open loop or PID control.