An iterative complex least squares algorithm for frequency-domain parametric system identification has been proposed. Extensive simulation results show that this algorithm converges rapidly and attains very accurate results, even in the presence of large measurement errors. This algorithm is also robust in the presence of over-ordering effects due to parasitics, so it is possible to identify a high-order physical system with a low-order ideal model. Thus, this algorithm can be simply applied to various tuning algorithms.
展开▼
