Improved Tag-based Indoor Localization of UAVs Using Extended Kalman Filter

Indoor localization and navigation of unmanned aerial vehicles (UAVs) is a critical function for autonomous flight and automated visual inspection of construction elements in continuously changing construction environments. The key challenge for indoor localization and navigation is that the global positioning system (GPS) signal is not sufficiently reliable for state estimation. Having used the AprilTag markers for indoor localization, we showed a proof-of-concept that a camera-equipped UAV can be localized in a GPS-denied environment; however, the accuracy of the localization was inadequate in some situations. This study presents the implementation and performance assessment of an Extended Kalman Filter (EKF) for improving the estimation process of a previously developed indoor localization framework using AprilTag markers. An experimental set up is used to assess the performance of the updated estimation process in comparison to the previous state estimation method and the ground truth data. Results show that the state estimation and indoor localization are improved substantially using the EKF. To have a more robust estimation, we extract and fuse data from multiple tags. The framework can now be tested in real-world environments given that our continuous localization is sufficiently robust and reliable.