Reuse-oriented SLAM Framework using Software Product Lines

Abstract

Simultaneous Localization and Mapping (SLAM) is a widely investigated problem in robotics. It depicts the process of a robot creating a map of an unknown environment while concurrently estimating its location within the self-created map. In recent years, many solutions have been proposed to the SLAM problem based on numerous approaches such as probabilistic filters or graph optimization. This work recognizes that, with the growing complexity and the active development in the field of SLAM, reuse is becoming an essential quality as researchers often have to solve architectural issues that are secondary to the core of the problem, which leads to sub-optimal realizations in the final SLAM product from the software point of view. Therefore, a component-based framework is introduced that regards reusability as a primary requirement of SLAM software, which highlights the core separable modules and implements them as encapsulated interchangeable components forming a software product line. The reusability of the framework is evaluated according to the reuse-readiness levels criteria and the results show improved modularity and reduction in the development and deployment time and effort.