Map Fusion in an Independent Multi-robot Approach

This paper concentrates on the study on the map fusion problem in the context of a multi-robot map building approach. Concretely it is seen as one of the steps towards the independent multi-robot map building. In the situation proposed a set of several robots performs map building tasks without the notion of other robots’ existence. Each robot builds its own local map using its observations and estimates its path independently. As a result, there will be a set of local maps that can be fused into a global one. This is the case when the map fusion takes importance. Particularly, we focus our experiments on landmark-based maps constructed using visual information and by means of a particle filter. When fusing two maps, we consider the uncertainty of the landmarks integrated by each different robot to its map. Key–Words: SLAM, map fusion, visual landmarks, particle filter, map alignment