Localization and map building based on particle filter and unscented Kalman Filter for an AUV

Simultaneous localization and mapping (SLAM) is of prime importance for navigation problem of autonomous underwater vehicle. Currently EKF-based SLAM and particle filter-based SLAM are prevalent methods though they have their own deficiency respectively. In this paper a modified RBPF method is proposed to apply in navigation and localization for our underwater vehicle, C-RANGER. Unscented Kalman filter instead of extended Kalman filter is used to incorporate the current observations as well as the historical observations into the proposal distribution. The simulation results show that the improved algorithm is more accurate and reliable while it is used to estimate the pose of AUV and locations of features.