Pattern Recognition in Real Time using Neural Networks: An Application for Pressure Measurement

Abstract

Retrieving information in real time from fringe patterns is a topic of great importance in scientific and engineering applications of optical methods. This paper describes an application of neural networks for real time pressure measurement using fringe pattern recognition. It is based on the capability of neural networks to recognize signals that are similar but not identical to the signals which were used to train the network. In this investigation a pressure sensor, which was part of the wall of the wind tunnel, and an optical apparatus were used to produce moire fringes. The fringe patterns generated were analyzed by a back propagation neural network at the speed of the recording device, which was a CCD camera. This method of information retrieval was used to measure the pressure fluctuations in the boundary layer flow. A second neural network was used to recognize the pressure patterns and to provide input to a control system that was capable to preserve the stability of the flow.