Multiple relative pose graphs for cooperative mapping

Abstract

This thesis describes a new representation and algorithm for cooperative and persistent simultaneous localization and mapping (SLAM) using multiple robots. Recent pose graph representations have proven very successful for single robot mapping and localization. Among these methods, iSAM (incremental smoothing and mapping) gives an exact incremental solution to the SLAM problem by solving a full nonlinear optimization problem in real-time. In this paper, we present a novel extension to iSAM to facilitate multi-robot mapping based on multiple pose graphs. Our main contribution is a relative formulation of the relationship between multiple pose graphs. Our formulation avoids the initialization problem and leads to an efficient solution when compared to a completely global solution. Efficient access to covariances at any time for relative parameters is also provided, facilitating data association and loop closing. Each individual pose graph still uses a global parameterization, so that the overall system provides a globally consistent multi-robot solution. The performance of the technique is illustrated on a publicly available multi-robot data set as well as other data including a helicopter-ground robot combination.