Experimental Evaluation of Indoor Drone Positioning Using Wi-Fi RTT and Machine Learning

When drones fly autonomously outdoors, they typically use GPS to implement navigation and safety functions. However, GPS is unstable and cannot satisfactorily provide these functions indoors. We have investigated indoor drone positioning based on Wi-Fi RTT (Round Trip Time). In this paper, we propose a method to estimate the position coordinate of a drone using machine learning. The proposed method uses pre-measured Wi-Fi RTT ranging results as training data to estimate the position coordinates of the stationary or flying drone. We compare and evaluate the estimation results of the machine learning-based method with those of the minimum mean square error-based method. The experimental results show that the machine learning-based estimation achieves more accurate results than the minimum mean square error-based estimation.