The vision based docking algorithm described here was motivated by, and implemented on, the ENDURANCE underwater autonomous vehicle for its four-week scientific mission to explore West Lake Bonney, Antarctica in December 2008. The vehicle entered the lake via a three meter vertical hole melted through the surface ice that was only slightly larger than the vehicle's diameter. During its daily scientific missions, ENDURANCE traveled a radial distance of up to 700 meters before returning to the melt-hole and surfacing. The vehicle's dead-reckoning navigation system was sufficient to bring it within a few of meters of the melt-hole, but it was not accurate enough to position the vehicle for safe egress. We therefore developed a homing method for centering the vehicle in the melt-hole and allowing it to rise to the surface while avoiding contact with the melt-hole walls.

PDF

In International Symposium on Unmanned Untethered Submersible Technology (UUST), 2009.

@inproceedings{murarka:uust09, title={Vision based frozen surface egress: A docking algorithm for the ENDURANCE AUV}, author={Aniket Murarka and Gregory Kuhlmann and Shilpa Gulati and Mohan Sridharan and Chris Flesher and Bill Stone}, booktitle={International Symposium on Unmanned Untethered Submersible Technology (UUST)}, url="http://www.cs.utexas.edu/users/ai-lab/?murarka:uust09", year={2009} }