An Acoustic Indoor Localization Method Based on Directional Variability for Mobile Robot

Abstract

Indoor localization is a critical requirement for mobile robots, and this paper proposes a cost-effective, accurate, and wide-ranging acoustic localization method. The localization system leverages acoustic microstructures to generate direction-specific information and consists of two components: anchors and a reader. A anchor consists of a carefully designed 3D-printed metamaterial structure called meta-cover and a speaker. The meta-cover cover the speaker and utilize acoustic diffraction phenomenon to embed the direction-specific signature in the emitted sound signals. The anchors are fixed in known positions and continuously emit sound signals. After a one-time calibration in the deployment environment, the system learns the directional signatures. The reader, fixed in the mobile robot, records the sound signals from multiple anchors and computes the position. This proposed method demonstrates excellent performance, with an average error of only 3° in direction estimation and an average localization accuracy of 6cm for mobile robot's localization.