Search papers, labs, and topics across Lattice.
This paper introduces a vision-in-the-loop framework for validating deep monocular pose estimation in autonomous UAV operations on maritime vessels, addressing the challenges of costly and risky at-sea validation. By integrating photorealistic rendered environments with onboard processing of a transformer-based pose estimator and employing a delayed Kalman filter for fusing vision and IMU data, the system effectively manages critical embedded effects like perception latency. Experimental results show stable closed-loop flight during autonomous takeoff, trajectory tracking, and landing, establishing a safer intermediate validation stage for maritime UAV autonomy before real-world deployment.
Hardware-validated simulations reveal that UAVs can achieve stable autonomous flight in complex maritime environments, bridging the gap between simulation and real-world deployment.
Autonomous UAV operations on ships require reliable vision-based relative pose estimation, yet at-sea validation is costly, weather-dependent, and risky. This paper presents a hardware-validated vision-in-the-loop framework that enables fully autonomous indoor flight while emulating photorealistic maritime environments. Rendered maritime views are processed onboard by a deep transformer-based monocular pose estimator. Delayed vision measurements are fused with high-rate IMU data using a delayed Kalman filter to provide consistent state estimates for geometric control. The system captures critical embedded effects, including perception latency, asynchronous updates, and computational constraints, that are absent in pure simulation. Autonomous takeoff, trajectory tracking, and landing experiments demonstrate stable closed-loop flight. The results establish a safe and hardware-realistic intermediate stage for developing maritime UAV autonomy prior to shipboard deployment.
