Abstract. Smart structures are a challenge for control system engineers due to the high number of distributed sensors and actuators needed. The control algorithms for these distributed and hierarchical control problems are currently not very well understood and new control methods have still to be developed. Neural network controllers possess the potential to learn from past examples and are thus a promising approach to the intended future intelligent control of smart structures. Using an indirect neural controller concept approach, a plant model is needed. In this paper we investigate the use of a dimensionally homogeneous neural network for nonlinear system identification and show the performance comparison to a typical neural identification technique.

Keywords: Neural Control, Smart Structures, Dimensional Analysis