Interval-based Sound Source Mapping for Mobile Robots

Axel Rauschenberger, Bernardo Wagner

Abstract

Auditory information can expand the knowledge of the environment of a mobile robot. Therefore, assigning sound sources to a global map is an important task. In this paper, we first form a relationship between the microphone positions and auditory features extracted from the microphone signals to describe the 3D position of multiple static sound sources. Next, we form a Constraint Satisfaction Problem (CSP), which links all observations from different measurement positions. Classical approaches approximate these non-linear system of equations and require a good initial guess. In contrast, in this work, we solve these equations by using interval analysis in less computational effort. This enables the calculation being performed on the hardware of a robot at run time. Next, we extend the approach to model uncertainties of the microphone positions and the auditory features extracted by the microphones making the approach more robust in real applications. Last, we demonstrate the functionality of our approach by using simulated and real data.

Download


Paper Citation