Higher order sliding modes observation, identification, and fault detection

The introduction in higher order sliding modes and higher order sliding mode differentiation is given. Algorithms for observation, identification and fault detection of linear time-invariant strongly observable systems with unknown inputs are developed, based on high order sliding mode differentiation. The possibilities of their extension to strongly detectable linear systems and nonlinear systems are discussed. The value of the equivalent output injection is used to identify perturbation directly. Continuous time versions of least square methods are proposed to identify unknown time variant and time-invariant parameters. Some applications of the proposed algorithms to of vehicle estimation parameters, backlash identification are presented actuator and sensors fault detection are discussed.