Titan SURF/Adaptive Modeling and Learning of Wind Fields

From DGC

Return to Titan SURF Home

Mentor(s): Alberto Elfes, Claire Newman

Project Description: A Titan aerobot will have to use knowledge about the wind field to compute trajectories that will allow it to reach high-value science targets on the surface of Titan. Current Titan wind models are characterized by high uncertainty, requiring an aerobot to conduct frequent measurements of the actual wind field during the execution of its mission. These measurements will have to be used to update the Titan wind field model and allow adaptive replanning of aerobot flight plans. The student will use an existing Titan wind model and develop algorithms to adaptively update (“learn”) these models using wind measurements obtained by a flying aerobot. These algorithms will be tested in an existing Titan mission simulation system. Depending on time, the algorithms may also be tested on an Earth wind model and actual data from the JPL aerobot field tests.

Recommended Prerequisites: Some basic experience with estimation and control, as well as experience with algorithm development.