Karsten Bohlmann and Andreas Beck-Greinwald and Sebastian Buck and Henrik Marks and Andreas Zell

Autonomous Person Following with 3D LIDAR in Outdoor Environments

1st International Workshop on Perception for Mobile Robots Autonomy (PEMRA 2012), Poznan, Poland, 2012


Abstract

The capability of a robot to follow autonomously a person highly enhances its usability when humans and robots collaborate. In this paper we present a system for autonomous following of a walking person in outdoor environments while avoiding static and dynamic obstacles. The principal sensor is a 3D LIDAR with a resolution of 59x29 points. We present a combination of 3D features, motion detection and tracking with a sampling Bayesian filter which results in reliable person detection for a low-resolution 3D-LIDAR. The method is implemented on an outdoor robot with car-like steering, which incorporates the target's path into its own path planning around local obstacles. Experiments in outdoor areas validate the approach.


Downloads and Links

[pdf]


BibTeX

@inproceedings{bohlmann2012pemra,
  author = {Karsten Bohlmann and Andreas Beck-Greinwald and Sebastian Buck and
	Henrik Marks and Andreas Zell },
  title = { Autonomous Person Following with 3D LIDAR in Outdoor Environments},
  booktitle = {1st International Workshop on Perception for Mobile Robots Autonomy
	(PEMRA 2012)},
  year = {2012},
  address = {Poznan, Poland},
  month = sep,
  abstract = {The capability of a robot to follow autonomously a person highly enhances
	its usability when humans and robots collaborate. In this paper we
	present a system for autonomous following of a walking person in
	outdoor environments while avoiding static and dynamic obstacles.
	The principal sensor is a 3D LIDAR with a resolution of 59x29 points.
	We present a combination of 3D features, motion detection and tracking
	with a sampling Bayesian filter which results in reliable person
	detection for a low-resolution 3D-LIDAR. The method is implemented
	on an outdoor robot with car-like steering, which incorporates the
	target's path into its own path planning around local obstacles.
	Experiments in outdoor areas validate the approach.},
  affiliation = {Chair of Cognitive Systems, University of Tuebingen, Department of
	Computer Science, Sand 1, 72076 Tuebingen, Germany},
  url = {http://www.cogsys.cs.uni-tuebingen.de/publikationen/2012/bohlmann2012pemra.pdf}
}