Landing a UAV in Harsh Winds and Turbulent Open Waters

TL;DR

This work tackles autonomous UAV landing on a USV in harsh open-water by introducing a nonlinear estimator-based MPC (MPC-NE) that operates without inter-vehicle communication. The framework combines a predictive USV model (FFT-based motion decomposition, Kalman observer, and wave forecasting) with a discrete UAV predictor, all within a real-time budget, and uses a novel objective function that includes a barrier-like landing term to achieve near-zero deck tilt. Key contributions include online USV motion observation, a sigmoid-based landing strategy, and demonstrated improvements over state-of-the-art methods in both simulations (including Gazebo) and real-world tests, achieving high landing success rates under wind up to 12 m/s and waves up to 4 m. The approach offers a robust, decentralized solution for UAV–USV cooperation in demanding maritime environments, with practical impact on autonomous offshore operations and emergency response.

Abstract

Landing an unmanned aerial vehicle unmanned aerial vehicle (UAV) on top of an unmanned surface vehicle (USV) in harsh open waters is a challenging problem, owing to forces that can damage the UAV due to a severe roll and/or pitch angle of the USV during touchdown. To tackle this, we propose a novel model predictive control (MPC) approach enabling a UAV to land autonomously on a USV in these harsh conditions. The MPC employs a novel objective function and an online decomposition of the oscillatory motion of the vessel to predict, attempt, and accomplish the landing during near-zero tilt of the landing platform. The nonlinear prediction of the motion of the vessel is performed using visual data from an onboard camera. Therefore, the system does not require any communication with the USV or a control station. The proposed method was analyzed in numerous robotics simulations in harsh and extreme conditions and further validated in various real-world scenarios.

Figures (7)

• Figure 1: uav landing on usv in real-world experiments.
• Figure 2: The mpc landing controller (yellow block) is integrated into the MRS system (grey blocks) and supplies the desired reference (velocity $\mathbf{\dot{r}}_d=\dot{x}\dot{y}\dot{z}^T$ and heading rate $\dot{\eta}_d$). In the MRS system, the first layer containing a Reference tracker processes the desired reference and gives a full-state reference $\bm{\chi}$ to the attitude controller. The feedback Position/Attitude controller produces the desired thrust and angular velocities ($T_d$, $\bm{\omega}_d$) for the Pixhawk flight controller (Attitude rate controller). The State estimator fuses data from Odometry & localization methods to create an estimate of the uav translation and rotation ($\mathbf{x}$, $\mathbf{R}$). Finally, the Vision-based Detector obtains the visual data from the camera and sends the pose information b of the usv to the mpc.
• Figure 3: An example illustration of the effective cost function values obtained during the landing approach.
• Figure 4: Comparison between the predictions made by the system using the onboard imu data (left) and using the vision data (right).
• Figure 5: Histogram comparison between the proposed approach and the standard approach during the touchdown of the uav on the usv deck.