Multiple direction-of-arrival estimation for a mobile robotic platform with small hardware setup

Abstract

Knowledge of how many users are there in the environment, and where they are located is essential for natural and efficient Human-Robot Interaction (HRI). However, carrying out the estimation of multiple Directions-of-Arrival (multi-DOA) on a mobile robotic platform involves a greater challenge as the mobility of the service robot needs to be considered when proposing a solution. This needs to strike a balance with the performance of the DOA estimation, specifically the amount of users the system can detect, which is usually limited by the amount of microphones used. In this contribution, an appropriately carriable small and lightweight hardware system (based on a 3-microphone triangular system) is used, and a fast multi-DOA estimator is proposed that is able to estimate more users than the number of microphones employed. © 2014 Springer Science+Business Media Dordrecht.