Enhancing processing time for graph-based SLAM applications

Simultaneous Localization and Mapping (SLAM) is the process that allows for a robot moving in unknown environment to build the map of the environment while simultaneously use this map to localize itself. Many approaches exist to solve this problem. Graph-based SLAM methods formulate the SLAM problem as a graph where the nodes represent robot and landmarks positions, and edges represent spatial constraints between nodes. In this paper, we present an optimized implementation of the incremental graph-based SLAM discussing the provided improvements. This implementation takes advantage of an optimized data structure and memory access to solve the nonlinear least squares problem related to the algorithm. To evaluate our implementation, we will evaluate the processing time of the implemented algorithm compared to those of the well known framework g2o.