Finite-Sample Analysis of Identification of Switched Linear Systems With Arbitrary or Restricted Switching

For the identification of switched systems with measured states and a switching signal, this letter aims to analyze effect strategies on estimation error. The data is assumed be collected from globally asymptotically or marginally stable under switches that are arbitrary subject an average dwell time constraint. Then system estimated by least-squares (LS) estimator. To capture parameters LS error, finite-sample error bounds developed in letter. obtained show logarithmic when there only modes; however, unstable modes, bound can increase linearly as parameter changes. This suggests presence strategy should properly designed avoid significant