Control relevant system identification using orthonormal basis filters

Control relevant system identification deals with developing a dynamic model that can be used for design of a controller from experimental data. In this paper an iterative technique based on orthonormal basis filters is used to develop reduced complexity, control relevant models from input-output data that is corrupted with noise. It is also shown that a parsimonious general orthonormal basis filter (GOBF) model can be developed from a crude estimate of time constant with no prior knowledge of time delay. Unlike ARX, ARMAX and similar techniques the method presented does not require prior knowledge of the time delay. The identification technique is tested using simulated data and the result is found to be sufficiently accurate. ©2007 IEEE.