Title: Input load identification using a holographic neural network

Authors: Wakae Kozukue, Ichiro Hagiwara, Hideyuki Miyaji

Addresses: Department of Mechanical Engineering, Kanagawa Institute of Technology, 1030 Shimo-ogino Atsugi-shi, Kanagawa-ken 243-0292, Japan. ' Department of Mechanical Sciences and Engineering, Tokyo Institute of Technology, 2-12-1 Ohokayama Meguro-ku, Tokyo 152-8552, Japan. ' Department of Mechanical Engineering, Kanagawa Institute of Technology, 1030 Shimo-ogino Atsugi-shi, Kanagawa-ken 243-0292, Japan

Abstract: A machine generates and undergoes dynamic loads during its operation. These dynamic loads are the main source of vibration and noise. If the dynamic loads can be identified exactly, it will become possible to provide the data effective for the reduction of vibration and noise. However, by the method of identification of dynamic loads of the conventional multiple-input systems, noise has a large influence on accuracy. Thus, in this paper, an identification method based on a neural network is proposed for independent multiple-input loads, and the results of the simulation by using the conventional method and the neural network are shown and compared in detail.

Keywords: dynamic loads; input load identification; multiple input–output system; MIMO; holographic neural networks; plate; transient response; Wiener filtering theory; vibration reduction; noise reduction; simulation; machine structures; vehicle design.

DOI: 10.1504/IJVD.2007.012302

International Journal of Vehicle Design, 2007 Vol.43 No.1/2/3/4, pp.173 - 183

Published online: 04 Feb 2007 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article