Sampled-data design of fixed structure LQ controllers

A method to design an output feedback sampled-data controller with fixed structure that minimizes a continuous quadratic loss function, thus accounting for intersample behavior, is presented. The method involves three steps: 1) a continuous linear quadratic optimal control problem is formulated; 2) the discrete equivalent problem is found; and 3) the problem is reformulated as a parametric optimization problem which is solved using the loss increment technique. In the feedback we have used an averaging sampler to prefilter the measured signals, thus circumvent the problems of sampling white noise.