Worst-case sampling for sampled-data H∞ design

Worst-case sampling of continuous-time systems is studied. The procedure generates a finite-dimensional discrete system with the property that for any discrete control law, the L₂-induced norm of the sampled-data system is less than a specified bound if and only if the H_∞ norm of the discrete system is less than the same bound. The worst-case sampling approach clarifies the connections between previous approaches to the sampled-data H_∞ problem, which were originally derived independently. The procedure also results in an efficient computational method for the sampled-data H_∞ control problem. As a particular application of the worst-case sampling approach, the solution of the sampled-data H_∞ problem for time-delay systems is presented