Multi-Robot Cooperative Localization Based on the Gauss-Hermite Particle Filter

SHAO Jin-xin,WANG Ling,WEI Xing
DOI: https://doi.org/10.3969/j.issn.1007-130x.2007.06.035
2007-01-01
Abstract:In the field of multi-robot cooperative localization,it is necessary to model accurate dynamic equations and observation equations, which need nonlinearity and non-Gauss systems. Several nonlinearity algorithms are applied in this realm. There are mainly two kinds. One is the extended Kalman filter (EKF),which makes a local linearization to a nonlinear system, so the Kalman filter (KF) can be utilized indirectly to filter and estimate. The other is the sequential Monte Carlo method on point mass, which is also called the particle filter(PF). In this paper, we introduce an improved particle filter,namely the Gauss-Hermite particle filter(GHPF).We lay our stress on comparing the effect of these three algorithms, which are applied in the realm of multi-robot cooperative localization.
What problem does this paper attempt to address?