Identification of Complex Systems with the use of Interconnected Hammerstein Models

摘要: Abstract In the paper we consider problem of identification large-scale interconnected systems. Accurate models complex nets are needed especially for optimal control in production and transportation The specifics lay fact that individual elements cannot be disconnected excited by arbitrary input processes purposes. Moreover, structural interactions cause correlations between interaction signals. particular, any output random disturbances can transferred into other inputs. It leads to cross-correlation problems, very difficult from effective modelling point view. First attempts made 1980's were limited static linear blocks, practice results rather devoted dynamic systems working steady state. this generalize approach components, which both nonlinear. All blocks represented two-channel Hammerstein (used e.g. real heating processes). least squares estimate is applied identify unknown parameters a system. particular obtained singular value decomposition procedure. algorithm as whole illustrated simple simulation example.