Design of a 3D Indoor Localization System Enabling Augmented Reality TV Applications

This paper focuses on the design of a robust Real Time Locating Systems (RTLS) based on the Ultra-Wide Band (UWB) technology for Augmented Reality (AR) applications in TV studios that require artists and/or a presenter to be accurately localized. According to a UWB-based measurement campaign, carried out in a TV studio environment, ranging measurements are heavily affected by the human body interference. Indeed, lots of outliers are present as the UWB receiver may synchronize to reflected paths, which result to be much stronger than the direct one. As a consequence, range errors are very large. In this context, to improve the localization performance, we increased the redundancy of the RTLS by employing more than one tag to localize the artists on the TV scene. In particular, we have applied the Extended Kalman Filter (EKF) algorithm to work with two and three tags. Moreover, an outlier detection and correction procedure have been defined and adopted for the ranging phase. The resulting localization performance, based on real range measurements, shows that the EKF with two tags outperforms by 83.5% the one with single tag.