Incremental Recursive Least-Squares Identification For The Systems Under Poor Observation Condition

For the autoregressive (AR) signal under poor observation condition, a kind of incremental recursive least-square (IRLS) algorithm is presented based on the incremental observation equation. Applying the convergence of RLS algorithm, it can be easily proved that the obtained estimates of model parameters and noise variance converge to the corresponding true values with probability one, i.e. they are strongly consistent. A simulation example shows its effectiveness.