Reviewing Wind Measurement Approaches for Fixed-Wing Unmanned Aircraft

Alexander Rautenberg,Jens Bange,Andreas Platis,Norman Wildmann,Martin Graf

doi:10.3390/atmos9110422

Alexander Rautenberg, Jens Bange + Show 3 more

Open Access

https://doi.org/10.3390/atmos9110422

Copy DOI

Journal: Atmosphere	Publication Date: Oct 28, 2018
Citations: 36	License type: CC BY 4.0

Affiliation: University of Tübingen, German Aerospace Center

Abstract

One of the biggest challenges in probing the atmospheric boundary layer with small unmanned aerial vehicles is the turbulent 3D wind vector measurement. Several approaches have been developed to estimate the wind vector without using multi-hole flow probes. This study compares commonly used wind speed and direction estimation algorithms with the direct 3D wind vector measurement using multi-hole probes. This was done using the data of a fully equipped system and by applying several algorithms to the same data set. To cover as many aspects as possible, a wide range of meteorological conditions and common flight patterns were considered in this comparison. The results from the five-hole probe measurements were compared to the pitot tube algorithm, which only requires a pitot-static tube and a standard inertial navigation system measuring aircraft attitude (Euler angles), while the position is measured with global navigation satellite systems. Even less complex is the so-called no-flow-sensor algorithm, which only requires a global navigation satellite system to estimate wind speed and wind direction. These algorithms require temporal averaging. Two averaging periods were applied in order to see the influence and show the limitations of each algorithm. For a window of 4 min, both simplifications work well, especially with the pitot-static tube measurement. When reducing the averaging period to 1 min and thereby increasing the temporal resolution, it becomes evident that only circular flight patterns with full racetracks inside the averaging window are applicable for the no-flow-sensor algorithm and that the additional flow information from the pitot-static tube improves precision significantly.

Highlights

Atmospheric boundary layer (ABL) studies are increasingly complemented by in situ measurements using small unmanned aircraft systems [1,2,3,4,5,6,7,8]
This is a typical value for averaging in meteorology, where, on the one hand, full racetracks are included and, on the other hand, the performance resulting from data only having fractions of racetracks is addressed
Each plot contains the results for both averaging periods, M = 240 s and M = 60 s, enabling a quantified comparison for each experiment (BAO in Figure 14, SNT in Figure 15, HEL in Figure 16, and PFR in Figure 17) between the averaging windows as well as between the algorithms

Summary

Introduction

Atmospheric boundary layer (ABL) studies are increasingly complemented by in situ measurements using small unmanned aircraft systems (sUAS) [1,2,3,4,5,6,7,8]. The capabilities of sUAS for meteorological sampling range from mean values for wind, thermodynamics, species concentration, etc., to highly resolved turbulence measurements, and from an accurate and diverse but larger sensor payload, down to small aircraft that can be operated from almost anywhere, with minimal logistical overhead. The common method for measuring the 3D wind vector from research aircraft is a multi-hole probe in combination with the measured attitude, position, and velocity of the aircraft. The attitude is measured with an inertial measurement unit (IMU), position, and velocity of the aircraft using a global navigation satellite system (GNSS) The combination of both systems, usually supplemented by an extended Kalman filter (EKF), is called an inertial navigation system (INS)

Methods

Results

Conclusion