”Towards informative path planning for acoustic SLAM”
C. Evers , A. H. Moore, and P. A. Naylor
Acoustic scene mapping is a challenging task as micro- phone arrays can often localize sound sources only in terms of their directions. Spatial diversity can be ex- ploited constructively to infer source-sensor range when using microphone arrays installed on moving platforms, such as robots. As the absolute location of a moving ro- bot is often unknown in practice, Acoustic Simultaneous Localization And Mapping (a-SLAM) is required in or- der to localize the moving robot’s positions and jointly map the sound sources. Using a novel a-SLAM approach, this paper investigates the impact of the choice of robot paths on source mapping accuracy. Simulation results de- monstrate that a-SLAM performance can be improved by informatively planning robot paths.