Optimal inputs in guaranteed identification problems

The problem of optimal choice of inputs for identifying control system parameters by measurement results is considered. A priori information is restricted by the knowledge of admissible limits of the variation of uncertain parameters and measurement errors. There is no statistical information; the identification problem is investigated within the framework of the minimax (guaranteed) approach. The optimal choice of inputs must provide the best quality of identification; the integral of the information function of the system is considered as an optimality criterion.