EECE 574 - Adaptive Control - Basics of System Identification
EECE 574 - Adaptive Control - Basics of System Identification
EECE 574 - Adaptive Control - Basics of System Identification
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Least-Squares <strong>Identification</strong><br />
Properties <strong>of</strong> the Least-Squares Estimates<br />
Note that the second term in the RHS is generally non-zero except when<br />
1 {w(t)} is zero mean and uncorrelated, or<br />
2 {w(t)} is independent <strong>of</strong> {u(t)} and y does not appear in the regressor<br />
FIR and step models<br />
Laguerre model<br />
Output-error models<br />
Only in the above cases is the least-squares estimate unbiased.<br />
Furthermore, if it is the case as the number <strong>of</strong> observations increases to<br />
infinity, then we say that the method gives consistent estimates.<br />
Guy Dumont (UBC) <strong>EECE</strong><strong>574</strong> - <strong>Basics</strong> <strong>of</strong> <strong>System</strong> <strong>Identification</strong> 22 / 40