There are many options for measuring local wind velocity and turbulence, but most are either immovable, cumbersome, or expensive. In search of a more affordable and flexible option, several works have turned their attention to unmanned aerial vehicles (UAVs), since an airborne wind-measuring system can provide the desired flexibility which makes it ideal for certain applications. The proposed method could be useful in complex terrain where accessibility with a traditional anemometer is challenging, as well as in situations where affordability is an issue. For example, wind conditions could be locally mapped in detail for design purposes of buildings, bridges, or other infrastructures, or to model and predict the accumulation of snow in cities. While many researchers use fixed wing UAVs (, Reference [
1,
2,
3]), others explore the possibilities of mounting an anemometer onto a multirotor drone (e.g., Reference [
4,
5]). Some, however, try to use the quadrotor itself as the measurement unit. Neumann and Bartholmai [
6] were pioneers in estimating wind velocity using a quadrotor’s internal stabilization unit. By relating thrust, drag, and gravitational forces and conducting wind tunnel tests, they could estimate horizontal wind speeds. Palomaki et al. [
5] applied the same idea but without using wind tunnels. Wolf et al. [
4] also tried to correlate roll and pitch angles to wind speed and direction. Mazzatenta et al. [
7] attempted to improve the accuracy of velocity estimates from a quadrotor’s inertial measurement unit (IMU) data by comparing measurements to those using Particle Image Velocimetry. Marino et al. [
8] evaluated the possibility of using a measure of differential thrust for wind estimation, while Wang et al. [
9] achieved wind estimates by considering wind as an acceleration disturbance of the rotorcraft. Gonzalez-Rocha et al. [
10,
11,
12] based their wind-velocity estimation on aircraft motion models, and similar methods were applied by Müller et al. [
13], Sikkel et al. [
14], and Schiano et al. [
15]. A recent study by Perozzi et al. [
16] proposed a wind estimate using time-varying parameter estimation algorithms, together with a quadrotor’s IMU data.
Within the field of quadrotor control, the vehicle’s position or velocity is often estimated in order to improve the control algorithms. Allibert et al. [
17] proposed a method of estimating the air velocity of a quadrotor in the body-fixed reference frame by designing an observer based on both aerodynamic theory and accelerometer data.
This paper, which is based on the author’s master thesis [
18], utilizes those equations developed by Reference [
17] and sets them in another context. By using the proposed equations together with measurements from sensors aboard the quadrotor, the quadrotor’s air velocity can be found and then transformed to the inertial reference frame. By then applying the wind triangle, an estimate of three-dimensional wind velocity can be found using only measurements of the quadrotor’s built-in sensors. This makes it an elegant and effective method for analyzing local wind conditions which can be used for assessing for example the wind energy potential. The largest benefit of the proposed method is that it makes it much easier to scan the wind conditions for a larger area, i.e., to find high laminar flow and avoid turbulence. The quadrotor can easily measure spatially distributed wind speed, which would be much harder to achieve with fixed masts and traditional anemometers. Furthermore, installing a mast or anemometer comes with much higher costs than using an off-the-shelf quadrotor.
The remainder of the paper is organized as follows: In
Section 2, Method, the IRIS+ quadrotor is presented; aerodynamic equations, as well as the parameters and variables needed to solve them, are discussed; the experimental setup for a static thrust test and flight tests are explained; and the means of data processing are presented.
Section 3 presents the results from the static thrust test, which provides some necessary parameters. The determination of the drag coefficient is explained, and the wind measurements and estimates from the flight tests are presented. In
Section 4, the results from the flight tests are discussed and presented in comparison to values from the literature.
Section 5 provides the conclusion and the
Appendix A contains the nomenclature.