Multiobjective control design using an identified model

This paper presents a model-based H₂/H_inf control design, in which the fact that the model is derived from an experimental identification process has been taken into account. Since system identification and control design must be treated as mutually dependent, this paper proposes an iterative scheme to improve the performance properties of the controlled system systematically. This method provides a controller that combines the H₂ guadratic performance criterion for disturbance rejection with the H_inf performance criterion for maximum robustness against destabilizing uncertainties. In this scheme, the desired criterion can be achieved by tuning several design parameters: γ for balancing between H₂ and H_inf norms; μ for loop transfer recovery; selection of the model dimension; etc. This paper also demonstrates the iterative scheme for the active suspension design using the half-car model.