This paper further develops a new approach to optimal experiment design for dynamical systems by introducing and comparing different robust optimal experiment design criteria. Until recently, solutions to the optimal experiment design problem were paradoxical in so far that the optimal experiment depended on the system parameters, i.e. the optimal experiment depended on the very thing that the experiment was aimed at finding. In a recent paper, we proposed an alternative approach, robust in nature, to the experiment design problem which, inter-alia, allowed for the possibility that the true parameters lay in a convex set. Here we further explore this robust solution. In particular, we examine the role played by different design criteria.
2006 American Control Conference. Proceedings of the 2006 American Control Conference (Minneapolis, Minnesota 14-16 June, 2006 ) p. 1659-1664