Adaptive Neural Network Control of Engine RPM 2004-01-2680
Conventional fixed controllers in combination with adaptive neural networks provide a powerful controller architecture. By utilizing the existing controller designs and augmenting them with adaptive neural networks engineers may exploit the merits of both control approaches. By adding on an adaptive component to the existing controller the range of operating conditions is increased and robustness to system degradation is improved. One of the simplest neural network controllers is the adaptive linear combiner. In this paper the adaptive linear combiner is described and the controller architecture is applied to an engine rpm controller. Results are given.