Advanced Kalman Filter-based Wind Velocity Estimation using Autonomous Aerial Vehicles

Hiranya Udagedara, Mahdis Bisheban

Keywords: Wind estimation, UKF, EKF, UAV

TL;DR: This study investigates the applicability of EKF and UKF for quadrotor UAV-based wind velocity measurement, indicating that the UKF achieves greater accuracy, while the EKF operates faster.

Abstract: Wind velocity measurements are important in meteorological studies and in extreme situations like wildfires. Among existing methods for wind velocity measurement, the use of quadrotor UAVs is widely adopted due to their compact size, maneuverability, and low cost. The Extended Kalman Filter (EKF) is a commonly used filter in wind velocity estimations, which linearizes the nonlinear equations of motion of the quadrotor UAV. Alternatively, the Unscented Kalman Filter (UKF) uses the exact nonlinear equations of motion for the estimation process. However, the use of UKF for wind velocity estimation using quadrotor UAVs is yet to be explored. In this work, the performances of EKF and UKF in wind velocity estimation with a quadrotor UAV were compared. The quadrotor UAV was tested under constant and sinusoidal wind disturbances through numerical simulations, while it was hovering, moving in a straight-line trajectory, and a Lissajous trajectory. Fixed disturbances and zero-mean Gaussian noise were introduced as process and measurement noises. Results indicate that UKF outperforms EKF in accuracy, while EKF remains advantageous in computational speed, offering a trade-off depending on application needs. Hence, UKF can be used when high precision is demanded. When moderate accuracy is demanded, EKF is the faster alternative. Though this study only focuses on wind estimation, it highlights the potential of UKF applications in environmental sensing using a quadrotor UAV.

Submission Number: 19