SURG Spring 2012

Vision-Based Object Following with an Autonomous Quadrotor

Project Members

Julian Binder, Chris Burchhardt, James Church, Jeff Cooper, Priyanka Deo,
Graham Harvey, Dan Jacobs, Sean James, Peter McHale, Nicolas Mellis, Doci Mou,
Tom Mullins, Kaven Peng, Harrison Rose, Matthew Sebek, George Nick Stanley,
Thomas Wucherpfennig, Alex Zirbel.

Advisor

Sanjiv Singh

Abstract

Robots are becoming a common sight in every aspect of modern life. In aerial robotics,
helicopters and other flying platforms are perfectfor surveillance in harsh environments
where ground -based navigation is extremely difficult or impossible - such as in dangerous
military zones or disaster-stricken regions. Aerial robots are also useful for supporting
ground vehicles: they can use their increased visual range to scout ahead, extend the range
of communication with outside entities, or provide supplies and tactical assistance in
emergency situations. However, the challenge of visually tracking and keeping up with an
object on the ground is difficult enough that current implementations usually involve a
human controller for the vehicle. This research aims to make the tracking of a grounded
object both autonomous and robust, so that the robot can provide support for a ground
vehicle without human assistance.

Google Doc

Google Doc Budget

Quadrotor_SURG_2012_S.pdf (180 KB) Alex Zirbel, 10/20/2011 12:08 am