A real-time fast incremental SLAM method for indoor navigation

Recently, numerous efficient approaches to simultaneously localization and mapping (SLAM) based on ground robots have been proposed. However, we may encounter difficulties when applying those algorithms to those systems with higher real-time requirements, such as micro aerial vehicles (MAVs). This paper presents a fast and effective solution to SLAM problems, enables a quadrotor to autonomously explore unknown indoor environments. We propose a probabilistic approach to estimate the position based on scan matching algorithm. The estimation uncertainty is computed by a cost function. Furthermore, Bayesian method is used to update the occupancy probability grid map, which provides an effective way to solve sensor uncertainty. Experimental results carried out by using a laser range sensor on a quadrotor platform in indoor environment show that the incremental SLAM strategy has a superior performance.