Extraction of Impact of Wind Turbulence on RC Helicopters Using Machine Learning

This paper presents a new cost effective wireless telemetry system capable of estimating ambient air turbulence using RC helicopters. The proposed telemetry system correlates the RC helicopter’s flight dynamics with ship air wake patterns generated by cruising naval vessels. The telemetry system consists of two instrumentation units each equipped with aviation grade INS/IMU sensors to measure dynamics of the helicopter with respect to the concerned naval vessel. The presented telemetry system extracts ship air wake patterns by removing the helicopter dynamic effects from actual measurements. This paper presents a comprehensive comparison between popular machine learning algorithms in eliminating effects of pilot inputs from helicopter’s dynamics measurements. The system was tested on data collected in a wide range of wind conditions generated by modified YP676 naval training vessel in the Chesapeake Bay area over a period of more than a year.