Recursive State-Space Identification of Non-Uniformly Sampled-Data Systems Using QR Decomposition

A recursive least-squares (LS) state-space identification method based on the QR decomposition is proposed for non-uniformly sampled-data systems. Both cases of measuring all states and only the output(s) are considered for model identification. For the case of state measurement, a QR decomposition-based recursive LS (QRD-RLS) identification algorithm is given to estimate the state matrices. For the case of only output measurement, another identification algorithm is developed by combining the QRD-RLS approach with a hierarchical identification strategy. Both algorithms can guarantee fast convergence rate with low computation complexity. An illustrative example is shown to demonstrate the effectiveness of the proposed methods.