Lighthouse Positioning System: Dataset, Accuracy, and Precision for UAV Research
This work addresses the need for affordable, distributed localization in UAV swarms, offering a practical solution for robotics research, though it is incremental as it builds on existing Lighthouse hardware.
The paper tackled the problem of enabling distributed UAV swarming by developing an open-source tracking software for the Lighthouse system that runs onboard small UAVs in real-time, and provided a dataset for flight and ground truth use-cases, analyzing accuracy and precision across different scenarios.
The Lighthouse system was originally developed as tracking system for virtual reality applications. Due to its affordable price, it has also found attractive use-cases in robotics in the past. However, existing works frequently rely on the centralized official tracking software, which make the solution less attractive for UAV swarms. In this work, we consider an open-source tracking software that can run onboard small Unmanned Aerial Vehicles (UAVs) in real-time and enable distributed swarming algorithms. We provide a dataset specifically for the use cases i) flight; and ii) as ground truth for other commonly-used distributed swarming localization systems such as ultra-wideband. We then use this dataset to analyze both accuracy and precision of the Lighthouse system in different use-cases. To our knowledge, we are the first to compare two different Lighthouse hardware versions with a motion capture system and the first to analyze the accuracy using tracking software that runs onboard a microcontroller.