Probabilistic self-localization and mapping - An asynchronous multirate approach

A new approach has been developed for multirate fusion, probabilistic self-localization, and map building for mobile robots. The asynchronous multirate holds are hybrid systems generating continuous signals from discrete sequence of inputs. This method has been used to overcome the filtering problem of dynamic system with inputs and outputs sampled at a different rates. It acts as interfaces for signals with different sampling rates, providing signals properly adapted to the required sampling period. It also reduce extrapolation errors in comparison to the naive solution of keeping the last updated value. The preliminary ideas of multirate extended Kalman filter (EKF) and unscented Kalman filter (UKF) were successfully applied to the mobile robot localization problem and three-dimensional tracking problem. The multirate asynchronous structure has been applied to particles filters (PFs) in fast simultaneous localization and map building (FastSLAM),as well as can be used with many other filters such as EKF, UKF, or discrete-Bayes filters.