Peak-minimizing design for linear control systems with exogenous disturbances and structured matrix uncertainties

A major characteristic of transients in linear dynamic systems with non-zero initial conditions is the maximum deviation of the trajectory from zero, which has a direct engineering meaning. If the maximum deviation is large, the so-called peak effect occurs. This paper completes a series of research works devoted to the peak effect in linear control systems. We consider a linear control system with non-random bounded exogenous disturbances and system uncertainties. A regular approach is proposed to design a stabilizing static state-feedback control law that minimizes the peak effect. The approach is based on the technique of linear matrix inequalities and reduces the original problem to a parameterized semidefinite programming one, which can be easily solved numerically. The proposed approach can be extended to new classes of problems, in particular, to the case of output feedback using an observer or a dynamic controller.