Evaluation of GP-SLAM in real-world environments

In this work, an introduction to a novel laser-based 2D Simultaneous Localization and Mapping (SLAM) approach, named as GP-SLAM, is made. The proposed method is based on a new type of map representation utilizing a regionalized Gaussian process map reconstruction algorithm. Two real-time experiments are conducted in indoor environments in order to evaluate the performance of this approach in real-world environments. In both experiments, a mobile robot equipped with a 2D laser scanner obtained its trajectories and learn the map of the environments simultaneously. The proposed method is compared with the widely-used Gmapping method in the identical scenarios. The results of those experiments testify the efficiency, accuracy, robustness and capacity of our approach to meet the requirements of real-time tasks.