A Machine Learning Approach for Estimating Air Data Parameters of Small Fixed-Wing UAVs Using Distributed Pressure Sensors

Abstract

This paper presents a method for estimating the air data parameters for a small fixed-wing, unmanned aerial vehicle (UAV) using an arrangement of low-cost Micro-electromechinal systems (MEMS)-based pressure sensors embedded in the surface of the UAV. The pressure measurements are used in a machine learning (ML) model to estimate the angle of attack, sideslip angle, and airspeed. Two ML algorithms based on artificial neural networks (NNs) and linear regression (LR) are implemented, tested, and assessed using data collected from wind tunnel experiments and a flight test, and the results are compared to a benchmark flight test. Training the ML algorithms using wind tunnel data was found to introduce several potential error sources that need to be addressed in order to provide accurate estimation on the benchmark flight test, whereas training the algorithms using flight data provides lower estimation RMSE values. The performance of the NN structures has been found to slightly outperform the LR algorithms in estimation accuracy. Finally, results from using different sensor configurations and a pseudo-Reynolds number are presented in an effort to evaluate the influence of sensor number and placement on the accuracy of the method.