On line H/sub 2/, H/sub infinity / and pointwise uncertainty bound quantification in identification of restricted complexity models

The authors consider the problem of online uncertainty bound quantification in identification of restricted complexity models. The algorithms presented provide a hard and tight upper bound on the unknown model uncertainty in H/sub 2/, H/sub infinity /, and pointwise senses. The algorithms proposed are simple, online, and recursive. This allows robust control and adaptive identification to be combined.<
>