Capacitive indoor positioning and contact sensing for activity recognition in smart homes

  • Authors:

  • Miika Valtonen;Timo Vuorela;Lasse Kaila;Jukka Vanhala

  • Affiliations:

  • (Correspd. E-mail: miika.valtonen@tut.fi);-;-;Department of Electronics, Tampere University of Technology, P.O. Box 692, 33101 Tampere, Finland

  • Venue:
  • Journal of Ambient Intelligence and Smart Environments
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

In smart homes, unobtrusive monitoring of user position and activity are important but challenging tasks. With the current state of technology, this task is especially hard to carry out in private areas where video surveillance is considered undesirable or even offensive. Even though some alternative methods for passive and unobtrusive monitoring of people have been proposed in the past, we still do not have a simple method that could be used to measure user position and activities as a single practical solution. To fulfill this need, this paper presents a single privacy-preserving method to measure user position and activity which can easily be adapted to measure the subject's height and posture as well. The system proposed in this paper can locate a person at floor level and monitor the subject's interaction with common household items such as a bed, sofa, table or refrigerator. The measurement method is based on the conductivity of the human body and on capacitive coupling of low-frequency signals between electrodes embedded in the floor and the in the environment. A test system was built for the TUT Smart Home and was evaluated with multiple test subjects, including a two-week-long living test to show the system's potential in long-term monitoring applications. The results show that a standing person can be positioned to within either 7 or 11-cm accuracy at a 90% confidence level using 30 × 30-cm and 60 × 60-cm-sized transmitting floor electrodes, respectively. For people walking, the respective accuracies are 17 and 33 cm. According to the long-term test results, the interactions with the environment were detected accurately. All the test data from this long-term living test, including the person's position, contact with common household items as well as the user annotations, have been made public and are available for download.