The WeAllWalk Data Set
Data files
Jun 08, 2016 version files 152.10 MB
-
WeAllWalkv3.0.zip
152.10 MB
Jun 08, 2016 version files 321.34 MB
Abstract
This data set contains time series from inertial sensors carried by blind subjects walking through pre-determined routes In two buildings in the UCSC campus. The sensors include two MetaWear CPRO (obtaining a tri-axial accelerometer and gyroscope) and two iPhone 6 containing try-axial accelerometer, gyroscopes, and magnetometers.
This data set is described in the paper:
Flores, German, and Roberto Manduchi. "WeAllWalk: An Annotated Data Set of Inertial Sensor Time Series from Blind Walkers", Proceedings of the 18th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2016). 2016.
Methods
The data is timestamped and annotated. In particular, the paths taken by the blind subjects are subdivided into continuous non-overlapping intervals corresponding, for example, to straight segments or left/right turns. In addition, our annotations contain specific "features", such as when a blind subject bumped onto an obstacles or stopped momentarily. Intervals and features are specified by their start and end time. An XML schema (WeAllWalk.xsd) has been defined to format these annotations, and is contained in the data set.
Please see the README file for detailed explanation of the data set content