Oliver Amft's Publications

Home Projects Publications Events Contact


Sorted by Date

Recognition of user activity sequences using distributed event detection

Oliver Amft, Clemens Lombriser, Thomas Stiefmeier, and Gerhard Tröster. Recognition of user activity sequences using distributed event detection. In EuroSSC 2007: Proceedings of the 2nd European Conference on Smart Sensing and Context, pp. 126–141, Springer, October 2007.

Download

[PDF]261.5kB  

Abstract

We describe and evaluate a distributed architecture for the online recognition of user activity sequences. In a lower layer, simple heterogeneous atomic activities were recognised on multiple on-body and environmental sensor-detector nodes. The atomic activities were grouped in detection events, depending on the detector location. In a second layer, the recognition of composite activities was performed by an integrator. The approach minimises network communication by local activity aggregation at the detector nodes and transforms the temporal activity sequence into a spatial representation for simplified composite recognition. Metrics for a general description of the architecture are presented. We evaluated the architecture in a worker assembly scenario using 12 sensor-detector nodes. An overall recall and precision of 77% and 79% was achieved for 11 different composite activities. The architecture can be scaled in the number of sensor-detectors, activity events and sequences while being adequately quantified by the presented metrics.

BibTeX

@INPROCEEDINGS{Amft2007-P_EuroSSC,
  author = {Oliver Amft and Clemens Lombriser and Thomas Stiefmeier and Gerhard
	Tr\"oster},
  title = {Recognition of user activity sequences using distributed event detection},
  booktitle = {EuroSSC 2007: Proceedings of the 2nd European Conference on Smart
	Sensing and Context},
  year = {2007},
  volume = {4793},
  series = {Lecture Notes in Computer Science},
  pages = {126--141},
  month = {October},
  publisher = {Springer},
  abstract = {We describe and evaluate a distributed architecture for the online
	recognition of user activity sequences. In a lower layer, simple
	heterogeneous atomic activities were recognised on multiple on-body
	and environmental sensor-detector nodes. The atomic activities were
	grouped in detection events, depending on the detector location.
	In a second layer, the recognition of composite activities was performed
	by an integrator. The approach minimises network communication by
	local activity aggregation at the detector nodes and transforms the
	temporal activity sequence into a spatial representation for simplified
	composite
	recognition. Metrics for a general description of the architecture
	are presented. We evaluated the architecture in a worker assembly
	scenario using 12 sensor-detector nodes. An overall recall and precision
	of 77\% and 79\% was achieved for 11 different composite activities.
	The architecture can be scaled in the number of sensor-detectors,
	activity events and sequences while being adequately quantified by
	the presented metrics.},
  doi = {10.1007/978-3-540-75696-5_8},
  file = {Amft2007-P_EuroSSC.pdf:Amft2007-P_EuroSSC.pdf:PDF},
  keywords = {Event detection - activity recognition - distributed detectors - activity
	sensing - inertial sensors - wireless sensor networks - smart objects},
  owner = {oam},
  timestamp = {2007/10/09}
}

Generated by bib2html.pl on Wed Dec 01, 2010 23:40:25