Set-membership approach to experiment planning for parameter identification in static regression models

This paper describes an approach to optimal experiment planning based on the assumption that the set of feasible models (defined by a priori assumptions and measurements that have already been taken) is a general set in a function space. In particular, this set does not have to be limited to conventional, linear-in-parameter models. The experiment planning procedure selects the observation locations sequentially so that at each step a conventional, linear in unknown parameters estimating model provides an optimal approximation to the set of feasible models