Exact bounded-error continuous-time linear state estimator

Simon Rohou, Luc Jaulin
Published in Systems & Control Letters.

Abstract. This paper proposes an interval-based method for estimating the state of a linear continuous-time dynamical system. In this work, we assume that the measurements are provided at discrete times and that all errors are bounded. Interval analysis is used to propagate the interval uncertainties continuously over time. The resulting method is guaranteed to never lose any feasible solution and provides an optimal polygonal enclosure of the state trajectory. A reproducible example illustrates the principle of the method.

Keywords: linear system, state-space equations, interval analysis, tubes, time uncertainties, state estimator

C++ sources of the application

See the source file

Tubex library

See the Tubex project.