DOB-based Wind Estimation of A UAV Using Its Onboard Sensor

TL;DR

The paper addresses the challenge of estimating 3-D ambient wind vectors for UAVs using only onboard sensors. It introduces a nonlinear disturbance-observer (DOB) framework that estimates the external wrench $\mathbf{f_e}$ and maps it to the relative air vector, enabling wind reconstruction via wind-triangle synthesis in real time, even during dynamic flight. A wind barrel and a decoupled front-end/back-end architecture support high accuracy ($0.11$ m/s speed error, $2.8^\circ$ directional error) and a measurement range up to $10$ m/s, validated in wind tunnel and field tests, with dynamic filtering to handle noise. The results demonstrate robust, cross-platform wind estimation with potential impact on meteorological field studies and UAV-based wind sensing, while noting limitations at very low wind speeds and incomplete vertical wind modeling as areas for future work.

Abstract

Unmanned Aerial Vehicles (UAVs) play a crucial role in meteorological research, particularly in environmental wind field measurements. However, several challenges exist in current wind measurement methods using UAVs that need to be addressed. Firstly, the accuracy of measurement is low, and the measurement range is limited. Secondly, the algorithms employed lack robustness and adaptability across different UAV platforms. Thirdly, there are limited approaches available for wind estimation during dynamic flight. Finally, while horizontal plane measurements are feasible, vertical direction estimation is often missing. To tackle these challenges, we present and implement a comprehensive wind estimation algorithm. Our algorithm offers several key features, including the capability to estimate the 3-D wind vector, enabling wind estimation even during dynamic flight of the UAV. Furthermore, our algorithm exhibits adaptability across various UAV platforms. Experimental results in the wind tunnel validate the effectiveness of our algorithm, showcasing improvements such as wind speed accuracy of $0.11$ m/s and wind direction errors of less than $2.8^\circ$. Additionally, our approach extends the measurement range to $10$ m/s.

Figures (9)

• Figure 1: Wind tunnel experiments with a given wind speed, the UAV remains hovering to collect data. Video available at: https://youtu.be/QpbzR2NJULg
• Figure 2: The UAV platform and coordinate system. A wind barrel is installed under the UAV to increase the wind estimation accuracy.
• Figure 3: Principe of wind estimation: (A) The relative air flow generates a total external force $\bm{f_e}$, which leads to new states $\bm{\Lambda}$. The force estimator uses the real-time states $\bm{\Lambda}$ to acquire the estimated external force $\hat{\bm{f_e}}$. The air-wind model utilizes the ground vector $\dot{p}$ and $\hat{\bm{f_e}}$ to estimate the wind speed $\bm{V_w}$ and wind direction $\vartheta$. (B) Front-end module: The force estimator acquires and processes the total states $\bm{\Lambda}$ to obtain the external force $\hat{\bm{f_e}}$. (C) Back-end module: Build a 3-D force-air model using a pre-calibrated force-air vector dataset. By incorporating the known ground vector $\dot{p}$ and applying the principles of the wind triangle, we can ultimately determine the wind vector and extract the wind speed and direction.
• Figure 4: 'Revolution-Thrust' curve for each motor.
• Figure 5: The horizontal force-air model.