Abstract
By implementing a combination of an activity recognition with a map-supported particle filter we were able to significantly improve the positioning of our navigation system for blind people. The activity recognition recognizes walking forward or backward, or ascending or descending stairs. This knowledge is combined with knowledge from the maps, i.e. the location of stairs. Different implementations of the particle filter were evaluated regarding their ability to compensate for sensor drift. © 2012 Springer-Verlag.
Author supplied keywords
Cite
CITATION STYLE
Schmitz, B., Györkös, A., & Ertl, T. (2012). Combination of map-supported particle filters with activity recognition for blind navigation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7383 LNCS, pp. 529–535). https://doi.org/10.1007/978-3-642-31534-3_78
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
