The concept of sliding mode control [11, 27, 32] has been extended to the problem of the state estimation by an observer, both for linear [10, 32] and nonlinear systems [1, 8, 28, 34]. Using the same design theory as variable structure control laws, the observer trajectories are constrained to evolve after finite time on a suitable sliding manifold by the use of a discontinuous output injection (the sliding manifold is usually given by the difference between the observer and the system output). The sliding motion provides an estimation (asymptotically or in finite time) of the system state. Sliding mode observers have been shown to be efficient in many applications, such as in robotics [3, 21], electrical engineering [6, 33], chemical reactors [25] and fault detection [14, 18].