Set-membership non-linear observers with application to vehicle localisation

This communication proposes to re-consider the localization problem (i.e. the configuration estimation) of a moving car as the state estimation of a discrete-time non-linear process. The use of set-theoretic concepts for state estimation is presented under the form of the computation of the direct and reverse images of a set by a function using interval analysis. One simplification of the computation of the direct image is shown for a class of non-linear systems and enables to reduce the computational complexity of the algorithm. The considered application belongs to this class and the method is applied to experimental data.