DISCRETE-TIME ESTIMATORS WITH GUARANTEED PEAK-TO-PEAK PERFORMANCE

摘要

In this paper is considered discrete-time estimation with worst-case peak-to-peak gain as the performance measure. Both linear and a restricted class of nonlinear estimation problems are treated. In the linear case estimators are synthesized that minimize the *-norm, the best upper bound on the induced l_∞ norm that one can obtain by bounding the reachable set with inescapable ellipsoids. In the nonlinear case conditions are presented that are sufficient to ensure a given level of peak-to-peak performance.