“…Given an appropriate positive constantū, our goal is to design a control u that is valued in [−ū,ū] and that can be computed from the outputs (2) and that renders (1) input-to-state stable with respect to δ = (δ 1 , δ 2 ). Choosingū ∈ (0, L 3 ) allows us to avoid the saturation in (1). Since the state space for (1) is R 3 , this implies as a special case that when δ = 0, all solutions of (1) for all constant initial states x(0) ∈ R 3 will converge to 0 as t → ∞.…”