Hybrid map-based SLAM with Rao-Blackwellized particle filters

The problem of Simultaneous Localization and Mapping (SLAM) is defined as to build a consistent map of the surrounding environment while simultaneously determining the vehicle trajectory within that map. Recently, approaches using Rao-Blackwellized particle filters (RBPFs) present an effective way to handle this problem. By Rao-Blackwellization scheme, they cannot only improve the estimation precision, but also reduce the overall computational complexity. In this paper, we propose a hybrid map-based SLAM using RBPFs. We describe the environment by using a grid map and a feature map together rather than using only one of them. Based on both maps, we have formulated a new proposal distribution which improves the robustness and efficiency of the algorithm. This makes the uncertainty of the predicted vehicle position decrease drastically. Therefore, we can obtain the same performance with a lower number of particles and the algorithm is able to run in real-time. Moreover, our approach also presents an adaptive solution which is able to select the best model to produce next generation particles depending on environment structures. Experimental results on simulated data illustrate the capabilities of our algorithm to improve the performance of SLAM compared to traditional approaches.