Andreas Masselli and Richard Hanten and Andreas Zell

Robust Real-Time Detection of Multiple Balls on a Mobile Robot

European Conference on Mobile Robots (ECMR 2013), Barcelona, Catalonia, Spain, September, 2013, pp. 1-9


Abstract

In this paper, we compare three different marker based approaches for six degrees of freedom (6DOF) pose estimation, which can be used for position and attitude control of micro aerial vehicles (MAV). All methods are able to achieve real time pose estimation onboard without assistance of any external metric sensor. Since these methods can be used in various working environments, we compare their performance by carrying out experiments across two different platforms: An AscTec Hummingbird and a Pixhawk quadrocopter. We evaluate each method's accuracy by using an external tracking system and compare the methods with respect to their operating ranges and processing time. We finally compare each method's performance during autonomous takeoff, hovering and landing of a quadrocopter.


Downloads and Links

[pdf]


BibTeX

@inproceedings{Masselli2013ecmr,
  author = {Andreas Masselli and Richard Hanten and Andreas Zell},
  title = {{Robust Real-Time Detection of Multiple Balls on a Mobile Robot}},
  booktitle = {European Conference on Mobile Robots (ECMR 2013)},
  year = {2013},
  address = {Barcelona, Catalonia, Spain},
  month = {September},
  url = {http://www.cogsys.cs.uni-tuebingen.de/publikationen/2013/masselli2013ecmr.pdf}
}

@inproceedings{Masselli2013,
  author = {Andreas Masselli and Shaowu Yang and Karl Engelbert Wenzel and Andreas
	Zell},
  title = {A Cross-Platform Comparison of Visual Marker Based Approaches for
	Autonomous Flight of Quadrocopters},
  booktitle = {Proceedings of International Conference on Unmanned Aircraft Systems},
  year = {2013},
  pages = {1-9},
  address = {Atlanta, Georgia, USA},
  month = may,
  abstract = {In this paper, we compare three different marker based approaches
	for six degrees of freedom (6DOF) pose estimation, which can be used
	for position and attitude control of micro aerial vehicles (MAV).
	All methods are able to achieve real time pose estimation onboard
	without assistance of any external metric sensor. Since these methods
	can be used in various working environments, we compare their performance
	by carrying out experiments across two different platforms: An AscTec
	Hummingbird and a Pixhawk quadrocopter. We evaluate each method's
	accuracy by using an external tracking system and compare the methods
	with respect to their operating ranges and processing time. We finally
	compare each method's performance during autonomous takeoff, hovering
	and landing of a quadrocopter.},
  days = {28-31},
  url = {http://www.cogsys.cs.uni-tuebingen.de/publikationen/2013/masselli2013icuas.pdf}
}