Nonlinear Relative Pose Estimation for Autonomous Shipboard Landing

Abstract

Shipboard aerial operations represent extreme and dynamic environments for landing aircraft.Rough seas and turbulent conditions lead to aircraft and vessel motions that can be large and seemingly unpredictable. To commence a safe landing, the relative position and attitude of the ship’s deck must be known and monitored. These conditions present adverse and challenging situations which serve as the motivation for this thesis. In the work reported here, a method of estimating the relative pose necessary for autonomous tracking and landing, using dual quaternions, is presented.Two nonlinear estimators were developed in the dual quaternion framework to use a Plucker coordinate based lidar measurement model. Synthetic measurements were generated from independent point models representing the aircraft and the ship. The estimators were tested using the synthetic measurements, yielding insights into the methods while successfully tracking the relative pose.