This incorporates a real plant model not necessarily resembling the plant but whose actuation error, y −yˆ , (where yˆ is the estimate of y) is driven to small proportions by setting a relatively short observer settling time, requiring large eigenvalues yielding high correction loop gains.