Hide/Show Apps

A Novel Map Merging Methodology for Multi-Robot Systems

Topal, Sebahattin
Erkmen, İsmet
Erkmen, Aydan Müşerref
In this paper, we consider the problem of occupancy grid map merging which is an important issue especially for multi-robot exploration task in search and rescue environments. We present scale invariant feature transform based methodology for combining individual partial map of robot units acquired from different parts of the mission environment. Proposed approach is designed not only for structured work areas, but also it is designed for unstructured and complex environment such as wide collapsed buildings. The proposed approach handles the limitations of existing works in the literature such as; first, merged robots' maps not required being in the same scale, secondly little overlapped region area between partial maps is enough for good merging performance and finally unstructured partial environment maps can be merged efficiently. The simulation results support the potential of the proposed methodology for occupancy grid map merging task.