Advanced engine control systems require accurate dynamic models of the comb
ustion process, which are substantially nonlinear. This contribution presen
ts the application of fast neural net models for engine control design purp
oses. After briefly introducing a special local linear radial basis functio
n network (LOLIMOT) the process of building adequate dynamic engine models
is discussed in detail. These neuro-models are then integrated into an uppe
r-level emission optimization tool which calculates a cost function for exh
aust versus consumption/torque and determines optimal engine settings. A DS
P-based process computer system allows a fast application of the optimizati
on tool at the engine test stand. (C) 2000 Elsevier Science Ltd. All rights
reserved.