Fault detection and diagnosis of an electrohydrostatic actuator using a novel interacting multiple model approach

In this paper, a new type of interacting multiple model (IMM) is introduced for the purposes of fault detection and diagnosis. The standard IMM is combined with a relatively new filtering method referred to as the smooth variable structure filter (SVSΓ). The SVSF is a type of sliding mode estimator, formulated in a predictor-correct fashion. It keeps the estimated state close to the true trajectory, and creates a stable estimation process. The combined method, referred to as the SVSF-IMM, is applied to an electrohydrostatic actuator (EHA). The results of the experiment are compared with the common form of the IMM, which utilizes the popular Kalman filter (KΓ).