Edge alignment of polyester (PET) films is important for achieving product quality and processing speed in winding, coating, drying, and other processes. The edge alignment can be achieved by lateral deflection control, provided that the film tension and transport speed are even at desired values. This article aims to correct the lateral deflection of films by designing robust controllers to swivel the guiding rollers and to maintain even tension and speed at target levels. The self-tuning neuro-proportional integral derivative controller and adaptive high-gain output feedback controller are adopted to guide the lateral deflection so that the film aligns at the desired position. A control scheme, neuron controller by associative learning, is used for maintaining tension and speed control. These strategies are applied to a simplified PET film processing system. The experimental results demonstrate that in our setup, the control schemes can effectively alleviate not only the lateral deflection but also the tension and speed fluctuation at target levels. (c) 2007 Wiley Periodicals, Inc.