DocumentCode :
1440847
Title :
A signal processing framework based on dynamic neural networks with application to problems in adaptation, filtering, and classification
Author :
Feldkamp, Lee A. ; Puskorius, Gintaras V.
Author_Institution :
Res. Lab., Ford Motor Co., Dearborn, MI, USA
Volume :
86
Issue :
11
fYear :
1998
fDate :
11/1/1998 12:00:00 AM
Firstpage :
2259
Lastpage :
2277
Abstract :
We present a coherent neural net based framework for solving various signal processing problems. It relies on the assertion that time-lagged recurrent networks possess the necessary representational capabilities to act as universal approximators of nonlinear dynamical systems. This applies to system identification, time-series prediction, nonlinear filtering, adaptive filtering, and temporal pattern classification. We address the development of models of nonlinear dynamical systems, in the form of time-lagged recurrent neural nets, which can be used without further training. We employ a weight update procedure based on the extended Kalman filter (EKF). Against the tendency for a net to forget earlier learning as it processes new examples, we develop a technique called multistream training. We demonstrate our framework by applying it to 4 problems. First, we show that a single time-lagged recurrent net can be trained to produce excellent one-time-step predictions for two different time series and also to be robust to severe errors in the input sequence. Second, we model stably a complex system containing significant process noise. The remaining two problems are drawn from real-world automotive applications. One involves input-output modeling of the dynamic behavior of a catalyst-sensor system which is exposed to an operating engine´s exhaust stream, the other the real-time and continuous detection of engine misfire
Keywords :
adaptive estimation; filtering theory; identification; modelling; pattern classification; recurrent neural nets; signal processing; EKF; I/O modeling; catalyst-sensor system; complex system; dynamic neural networks; engine misfire; extended Kalman filter; input-output modeling; multistream training; nonlinear adaptive filtering; nonlinear dynamical systems; one-time-step predictions; process noise; recency; robustness; signal processing framework; system identification; temporal pattern classification; time series; time-lagged recurrent neural nets; time-series prediction; universal approximators; weight update procedure; Adaptive filters; Adaptive signal processing; Filtering; Neural networks; Nonlinear dynamical systems; Pattern classification; Recurrent neural networks; Signal processing; System identification; Vehicle dynamics;
fLanguage :
English
Journal_Title :
Proceedings of the IEEE
Publisher :
ieee
ISSN :
0018-9219
Type :
jour
DOI :
10.1109/5.726790
Filename :
726790
Link To Document :
بازگشت