Robust set‐membership state estimator against outliers in data

Abstract

Based on interval computation, a set-membership state estimator capable to manage a certain type of outliers in measurements is proposed for uncertain discrete-time linear systems. To achieve this purpose, two set-valued filtering techniques are implemented in the presented state estimation algorithm. The setting up of these techniques offers two main advantages. On one hand, the convergence of the estimated state enclosures width is guaranteed and, on the other hand, the algorithm robustness against outliers in data is ensured. That is, unlike former methods, the proposed set-valued state estimator preserves the framing property despite the presence of some false values in available sensors data. To show the efficiency and the performance of the introduced set-valued state estimator, it is compared, through two numerical examples, with an optimal interval observer selected from the literature for its high performance.