A note on sampled-data observers

Abstract We present a new design methodology of sampled-data observers that unifies and generalizes many existing design methodologies. The approach is based on continuous-time observers that feature two characteristics: (i) exhibit exponential convergence in the noiseless case with respect to a given Observer Lyapunov Function, and (ii) satisfy an Input-to-Output Stability property with respect to output measurement noise. The design approach applies to a wide class of systems and yields sampled-data observers that inherit these two performance characteristics of the underlying continuous-time observer. The main component of the proposed sampled-data observer is a novel output predictor that encompasses both inter-sample predictors and Zero-Order-Hold predictors.