Self-learning FNN (SLFNN) with optimal on-line tuning for water injection control in a turbo charged automobile

Abstract

This paper proposes a new architecture of self-learning fuzzy-neural-network (SLFNN) for water injection control in a turbo-charged automobile. The major advantage of SLFNN is that no off-line training is needed for initialization. The SLFNN will initialize itself with a random set of initial weighting factors (normally zeros) and a specifically designed on-line optimal training algorithm is invoked immediately after the engine of the automobile is turn on. The on-line optimal training can guarantee that the weighting factors will be directed toward a maximum-error-reduced direction. Although this SLFNN can also be used as a controller for fuel injection, we adopt the SLFNN as the water injection controller to reduce the knocking effects of a turbo-charged engine and therefore the emission is cleaner with less petrol consumption. Real implementation has been performed in a Saab NG 900 (1994 -1998) automobile with excellent results.