Online frequency domain system identification based on a virtual instrument

System identification is an inherently iterative process. Yet, limited attempts have been made so far to implement the whole identification loop in a single device. This paper discusses the difficulties of the task and presents a solution based on a Matlab toolbox and a set of virtual instruments. During the identification session, the obtained models may call for refinement or validation by new experiments. Using this integrated software-hardware tool, these subsequent experiments can be accomplished online with the identification procedure. As a demonstration, the modeling of a hairdryer is described. The universal applicability of this solution is believed to be guaranteed by the modular architecture of virtual instrumentation and the general definition of the software interface developed. The interface allows combining Matlab-based identification packages with virtual instruments or pure hardware interfaced to Matlab.