Collaboration and Interventions on Urban Environment Mapping with Graph-based SLAM Algorithm

The article describes the implementation of a collaborative approach for localization and mapping of large urban environments. This approach foresees the presence of multiple human operators, endowed with wearable sensors, whose outputs are processed in real-time in order to estimate human odometry. A Graph-based CLAM approach, generating a global map, is then applied: in order to improve the optimization phase of the algorithm, deterministic and probabilistic intervention approaches are proposed. More in detail, the loop closure procedure has been improved according to human sensing and the observations of external sensors. Experimental tests have been carried out in the historical center of Genova, proving that the proposed method overcomes classical multi-operator SLAM approaches.