A frequency domain approach for local module identification in dynamic networks

In classical approaches of dynamic network identification, in order to identify a system (module) embedded network, one has formulate Multi-input-Single-output (MISO) identification problem that requires parametric model for all the modules constituting MISO setup including (possibly) noise model, and determine their order. This requirement leads selection steps are no interest experimenter which increases computational complexity large-sized networks. Also, using (like BJ method) can suffer from local minima, however neglecting its impact on variance estimates. this paper, we provide two-step approach avoid these issues such is independent network. The first step involves performing non-parametric indirect get frequency response function (FRF) estimates as frequency. second step, estimated FRF target module smoothed domain estimator with previous model. developed practical weak assumptions noise, uses available toolbox, only interest, reduce Numerical simulations illustrate potentials introduced method comparison methods.