A passive isolation of sensor faults from un-stealthy attacks in uncertain nonlinear systems

In this paper, a novel nonlinear observer-based approach is established to cast light on the problem of isolating sensor faults from un-stealthy attacks for Lipschitz affine nonlinear systems subjected to unknown uncertainties and disturbances. A structure consisting of a multi-line transmission mechanism, a random number generator (RNG), and Dedicated Observer Scheme is developed to make the abnormal behaviors distinguishable. A group of two interlinked nonlinear Luenberger-Like observers is designed in nonlinear coordinates to detect abnormal behaviors. Two banks of detection observers and an incidence matrix are introduced to make the final decisions. Adaptive threshold techniques are exploited to guarantee the robustness against the model uncertainties and disturbances. Compared with the existing results, the proposed approach isolates the abnormal behaviors without the need for any redundant hardware components. Finally, the performance of the proposed approach is evaluated on a continuous stirred tank reactor (CSTR).