UAV-USV docking control system based on motion compensation deck and attitude prediction

• Published in: Ocean engineering 2024-09, Vol.307, p.118223, Article 118223
• Main Authors: Tian, Ennong, Li, Ye, Liao, Yulei, Cao, Jian
• Format: Article
• Language: English
• Subjects: Docking control; Motion compensation; Quiescent period prediction; Unmanned aerial vehicle; Unmanned surface vehicle
• ISSN: 0029-8018; 1873-5258
• DOI: 10.1016/j.oceaneng.2024.118223

Unmanned Aerial Vehicle (UAV) possess excellent maritime detection capabilities and have a fast cruising speed. Developing a collaborative mission system including UAV and Unmanned Surface Vehicle (USV) can significantly enhance the efficiency of maritime operations, making it a crucial developmental path for heterogeneous unmanned maritime systems. However, the rocking motion of USV due to sea wind and waves makes stable takeoff and landing of UAV difficult. To overcome this issue, we propose a pitch motion compensation deck that ensures the accuracy of sensor positioning and the stability of relative attitude during the docking process. To address the periodic changes in roll resulting from the interaction between the USV and waves, we present an online rolling prediction method based on the Hilbert-Huang Transform-Long Short-Term Memory network (HHT-LSTM). It segments the rolling time history data obtained from USV inertial navigation into distinct working conditions, which are fed into the trained LSTM network for training and forecasting. The optimal time for UAV landing is then determined using this data, which is sent to the UAV's motion controller to complete its precise landing. We validated the stability and feasibility of our proposed algorithm through practical experiments conducted on field water surfaces. •A low-cost pitch motion compensation flight deck mechanism is designed, which can adaptively and stably adjust the pitch angle according to the USV speed and the influence of waves. It not only ensures the accuracy of the UAV's UWB positioning data, but also avoids accidents in which the propeller hits the deck surface due to excessive trim when the UAV is docked with the USV.•By combining the HHT with LSTM, an LSTM network is selected based on the stability of the instantaneous frequency of the USV's roll history information. This selection allows for the prediction of either the roll value or the minimum value of the roll envelope, thus determining the optimal landing moment for the UAV. The calculated landing time is then transmitted to the motion planning controller of the UAV, enabling a stable UAV landing.•A prototype design that satisfies the experiment was proposed, and the feasibility of the proposed prototype and algorithm was verified through multiple field water surface experiments.