Abstract
This paper presents an algorithm for simultane- ous localization and mapping for a mobile robot using monocular vision and odometry. The approach uses Variable State Dimension Filtering (VSDF) flame- work to combine aspects of extended Kalman filtering (EKF) and nonlinear batch optimization. This pa- per describes two primary improvements to the VSDF. The first is to use the maximally informative statis- tics criterion to derive an interpolation scheme for lin- earization in recursive filtering. The interpolation is based on fitting a set of deterministic samples rather than using analytic Jacobians. The second improve- ment is to replace the inverse covariance matrix in the VSDF with its Cholesky factor to improve the compu- tational complexity. Results of applying the filter to the localization and mapping are presented.

This publication has 6 references indexed in Scilit: