Distributed Frobenius-Norm Finite Memory Interacting Multiple Model Estimation for Mobile Robot Localization

This paper proposes a novel state estimation algorithm, called the distributed Frobenius-norm finite memory interacting multiple model (DFFM-IMM) estimation algorithm, for mobile robot localization in wireless sensor networks (WSNs). The proposed algorithm involves finite memory estimation based on recent finite measurements; such estimation facilitates robust localization in cases of missing measurements and robot kidnapping. Furthermore, the proposed algorithm employs IMM, which facilitates accurate localization if a mobile robot abruptly changes its speed and course. Notably, average-consensus-based distributed processing renders the proposed DFFM-IMM algorithm computationally efficient, and hence, real-time processing for very short sampling times of the WSN is possible. The proposed algorithm’s performance is demonstrated by comparing it with a centralized Frobenius-norm finite memory IMM (CFFM-IMM) estimation algorithm and a localization algorithm on the basis of simulations and experiments.