Onboard Wind-Preview Model Predictive Control Using Pitot-Static Sensing for Multirotor UAVs

Bas Meere; Eline Wisse; Laurens Vousten; Sander Doodeman; Elena Torta; Paula Chanfreut; Duarte Antunes

cs.ROSep 25, 2026

Onboard Wind-Preview Model Predictive Control Using Pitot-Static Sensing for Multirotor UAVs

Authors: Bas Meere, Eline Wisse, Laurens Vousten, Sander Doodeman, Elena Torta, Paula Chanfreut, Duarte Antunes

Organizations: Department of Mechanical Engineering, Eindhoven University of Technology, 5612 AE Eindhoven, The Netherlands, in the Control Systems Technology group · Department of Mechanical Engineering, Eindhoven University of Technology, 5612 AE Eindhoven, The Netherlands, in the Robotics group

Abstract

Effective wind gust rejection and stable hovering are critical for the outdoor operation of autonomous drones. However, existing gust rejection methods are primarily reactive, inferring the disturbance from the resulting motion or measuring it at the airframe. Either way, the wind has already begun to act before it can be compensated. In this work, we anticipate the gust instead by measuring the wind ahead of the drone with a low-cost, low-weight pitot-static sensor mounted on a boom. The resulting wind preview is incorporated into a nonlinear model predictive controller (MPC), which optimizes the drone motion while anticipating wind disturbances. A longer boom offers more preview time but adds inertia and degrades flight performance. We characterize this trade-off in simulation and show that the optimal preview distance is not a fixed property of the platform, but shifts with the wind speed and with how quickly the drone can respond. Indoor hardware experiments confirm the trend and show that the proposed controller substantially improves hover performance against a PX4 baseline and an otherwise identical wind-unaware MPC. Outdoor experiments show that the error along the wind direction is reduced by 54 percent with respect to the baseline, demonstrating that a single wind-aligned sensor can significantly improve hovering performance.

Figures & tables

Fig. 1 : Outdoor flight experiment with a boom-mounted pitot-static sensor. The boom measures the wind upstream of the drone, giving the MPC a preview of incoming gusts to improve hover performance in windy conditions.

Fig. 2 : Angle sensitivity of the pitot-static sensor. A pressure coefficient $C_{p}$ = 1 means the probe registers the full flow. Two repeated sweeps are shown.

Fig. 3 : Wind speed measurement comparison between a static anemometer and the pitot tube mounted on a hovering drone. Even at low wind speed and while hovering the pitot sensor is able to capture the wind speed effectively.

Fig. 4 : Control diagram of the wind-preview MPC including wind sensing and the PX4 inner loop. The boom length

D

is shown dashed as it is a configuration parameter rather than a runtime signal.

Parameter	Value	Parameter	Value
$m_{\mathrm{drone}}$	2.06 kg	$Q$	diag(5,5,3,1,1,1,6,6,6)
$\mu_{\mathrm{boom}}$	41 g $\text{m}^{-1}$	$R$	diag(0.1,1,1,1)
$m_{\mathrm{sensor}}$	5 g	$P$	2 $Q$
$A$	0.1 $\text{m}^{2}$	$N$	30
$C_{d}$	0.5	$T_{\mathrm{s}}$	1/30 s
$\alpha_{\mathrm{EMA}}$	0.05	$T_{\mathrm{MA}}$	60 s (0.1 s indoor)

TABLE I : Parameters used in simulation and on hardware.

Fig. 5 : Segment of longitudinal, lateral and vertical wind components of a 7 m/s field generated under IEC class A normal turbulence. Note the different vertical scales.

Fig. 6 : Position error in $x$ for a simulated flight with three different boom lengths.

Fig. 7 : Standard deviation of the position error in $x$ against boom length, for four mean wind speeds. Longer booms give more preview time but add inertia, producing an optimum at 1.25 m for 4 and 7 m/s (circled). At 2 m/s the drone is fast enough that boom length has little effect, and at 10 m/s the preview time is too short to help much.

Fig. 8 : Performance improvement with boom length at 4 m/s wind for an artificially slowed drone, obtained by limiting the maximum body rate in the MPC (deg/s). Slower platforms benefit more from preview.

Fig. 9 : Indoor experimental setup where the drone hovers in front of a fan, with the boom-mounted pitot tube pointing into the flow. A sheet in front of the fan is opened rapidly to generate a repeatable gust.

Fig. 10 : (Top) Mean position error (solid) with one standard deviation (shaded) for each controller. RMSE over the window is 0.144 m for the PX4 baseline, 0.120 m for the wind-unaware MPC, and 0.048 m for the wind-aware MPC. (Bottom) Mean wind speed at the boom, pooled across all runs.

Fig. 11 : (Top) Mean position error (solid) with one standard deviation (shaded) for four different boom lengths. (Bottom) Mean wind speed at the boom, pooled across all runs.

Boom length [m]	RMSE [m] (2 m/s)	RMSE [m] (4 m/s)
0.5	0.0593 (0.010)	0.0836 (0.009)
0.75	0.0602 (0.008)	0.0722 (0.009)
1.0	0.0569 (0.010)	0.0652 (0.017)
1.25	0.0552 (0.006)	0.0538 (0.009)

TABLE II : Hardware performance for different boom lengths at two wind speeds. Values are the mean RMSE over 10 runs, with the standard deviation in parentheses.

Fig. 12 : Position error for the outdoor experiments. RMSE along

x

, the direction aligned with the boom, improves by 54 percent, while

y

stays similar. The setpoint changes to the current position when changing controllers.

Quadrotor Control

Quadrotor control studies how to stabilize, guide, and adapt aerial robots under demanding flight conditions and physical constraints.

See all 126 papers

Wind-Aware Reinforcement Learning Control of a Small Quadrotor Using Learned Onboard Wind Estimation in Simulated Atmospheric Turbulence
Jul 1, 2026Abdullah Al Tasim, Wei SunRL Control Aerial Robotics
University of Oklahoma, Norman, Oklahoma, 73019
WESPR: Wind-adaptive Energy-Efficient Safe Perception & Planning for Robust Flight with Quadrotors
Mar 10, 2026Khuzema Habib, Pranav Deshakulkarni Manjunath, Kasra Torshizi +2Quadrotor Control Motion Planning
University of Maryland, College Park.
GustPilot: A Hierarchical DRL-INDI Framework for Wind-Resilient Quadrotor Navigation
Mar 20, 2026Amir Atef Habel, Roohan Ahmed Khan, Fawad Mehboob +2UAV Navigation Quadrotor Control
Intelligent Space Robotics Laboratory, Center for Digital Engineering, Skolkovo Institute of Science and Technology, Moscow, Russia.

Jul 1, 2026J/K move · Enter open · S save

Wind-Aware Reinforcement Learning Control of a Small Quadrotor Using Learned Onboard Wind Estimation in Simulated Atmospheric Turbulence

Abdullah Al Tasim, Wei Sun

University of Oklahoma, Norman, Oklahoma, 73019

RL Control Aerial Robotics Quadrotor Control Atmospheric Turbulence Mitigation

Open paper AskPDF arXiv

Small multirotor aircraft are increasingly tasked with operations in the atmospheric boundary layer, where turbulent winds comparable to the vehicle's airspeed degrade trajectory tracking and can defeat conventional feedback control. This work illustrates a two-stage learning pipeline that first estimates the local wind from onboard kinematics and dynamics and then exploits that estimate inside a reinforcement learning (RL) flight controller. The wind estimator, an attention-augmented gated recurrent network trained on thousands of simulated flights through von Karman turbulence with power-law shear and veer, recovers the horizontal wind vector with a per-flight root-mean-square error of 0.40 m/s and a direction error of 3.2 degrees on unseen wind regimes, an accuracy near the floor imposed by unresolved turbulence, and generalizes to vertical ascent profiles with a skill score of 0.861 over a constant-wind reference. A proximal policy optimization controller receiving the frozen estimator's output reduces horizontal trajectory tracking error by 48% relative to a wind-blind proportional-derivative baseline across mean winds of 4 m/s to 12 m/s, winning on 100% of evaluation episodes. A three-way ablation decomposes this improvement into a kinematic component, available without wind information, and a wind-perception component; the perception share rises with wind speed, from small in light winds toward roughly half the total benefit in strong winds, consistent with the quadratic scaling of aerodynamic drag. The controller degrades gracefully on out-of-distribution winds of 13 m/s to 15 m/s, where the baseline fails catastrophically.

Abstract

Quadrotor Control

Explore similar work

Wind-Aware Reinforcement Learning Control of a Small Quadrotor Using Learned Onboard Wind Estimation in Simulated Atmospheric Turbulence

WESPR: Wind-adaptive Energy-Efficient Safe Perception & Planning for Robust Flight with Quadrotors

GustPilot: A Hierarchical DRL-INDI Framework for Wind-Resilient Quadrotor Navigation