Development of compact forward and inverse estimators by using neural networks Conference

Tansel, IN, Chen, P, Yenilmez, A et al. (2005). Development of compact forward and inverse estimators by using neural networks .

cited authors

  • Tansel, IN; Chen, P; Yenilmez, A; Lindsay, H; Vu, B

fiu authors

abstract

  • Backpropagation type neural networks were proposed to represent the forward and backward relationships between the sensory data and critical parameters. Their performance was evaluated by teaching them the relationship between the operating conditions and generated sound pressure levels (SPL). Several neural networks were used in the study to make forward and inverse estimations after they were trained by using the data obtained from the Rocket Acoustic Prediction Tool (RAPT). The average estimation error of the neural networks was less than 3% on simulated cases. Study indicated that neural networks can be used for development of smart sensors which evaluate acoustic loading and fatigue during the launching of rockets. Copyright © 2005 by the American Institute of Aeronautics and Astronautics, Inc.

publication date

  • December 1, 2005