Using Time Use with Mobile Sensor Data: A Road to Practical Mobile Activity Recognition?

Description

Having mobile devices that are capable of finding out what activity the user is doing, has been suggested as an attractive way to alleviate interaction with these platforms, and has been identified as a promising instrument in for instance medical monitoring. Although results of preliminary studies are promising, researchers tend to use high sampling rates in order to obtain adequate recognition rates with a variety of sensors. What is not fully examined yet, are ways to integrate into this the information that does not come from sensors, but lies in vast data bases such as time use surveys. We examine using such statistical information combined with mobile acceleration data to determine 11 activities. We show how sensor and time survey information can be merged, and we evaluate our approach on continuous day-and-night activity data from 17 different users over 14 days each, resulting in a data set of 228 days. We conclude with a series of observations, including the types of activities for which the use of statistical data has particular benefits.

In this project we use a strategy first suggested by Partridge and Golle [2] that utilizes time use diaries in an activity recognition process. We follow our work from [1], using specific time use survey features to infer the user's activity, but also to improve on the results from a common classifier.

The scripts used in this study visualize and evaluate the mobile sensor data obtained from 17 different participants.

Further details can be found in the paper.

Fig. 1: Depicting per time-of-day the percental occurrences of common activities over all male participants from the German Time Use Survey of 2001/2002.
 
 
[1] M. Borazio nad K. Van Laerhoven, "Improving Activity Recognition without Sensor Data: A Comparison Study of Time Use Surveys". AH'13
[2] K. Partridge and P. Golle, "On Using Existing Time-Use Study Data for Ubiquitous Computing Applications". UbiComp'08
 

Downloads

Download the database and the scripts.

Citation / first paper

Marko Borazio and Kristof Van Laerhoven, "Using Time Use with Mobile Sensor Data: A Road to Practical Mobile Activity Recognition?". 12th International Conference on Mobile and Ubiquitous Multimedia (MUM2013), ACM Press, Lulea, Sweden, 2013.

Important note

Disclaimer: This data is property of the Embedded Sensing Systems Group, FB Informatik, TU Darmstadt. You may use this data for scientific, non-commercial purposes, provided that you give credit to the owners when publishing any work based on this data.

alte website
print print | print print | Impressum impressum