KNOWME Network

Wireless Body Area Networks (WBANs) constitute a new class of wireless sensor networks that enable diverse applications in healthcare, entertainment, sports, military and emergency situations. A WBAN typically comprises of few biometric sensors (e.g. accelerometers, electrocardiographs), which measure vital signs such as heart-rate, oxygen levels, etc, and an energy-constrained personal device (e.g. cellphone, PDA). The KNOWME network is a prototype WBAN that integrates off-the-shelf sensors with a Nokia N95 mobile (see Figure 1 below) to continuously monitor and analyze the biometric signals of an individual, targeting obesity-related applications. It is the outcome of a multiyear collaboration between Electrical Engineering, Computer Science and Preventive Health researchers at the University of Southern California.

Practical realization of WBANs faces a number of unique challenges. Among these, the KNOWME network team has focused on 4 main research thrusts:

  • System architecure: develop solutions to ensure efficient operation, system viability and application stability.
  • Multimodal physical activity recognition: develop intelligent feature extraction and selection techniques and design accurate classifiers.
  • Energy efficient sensor selection: design communication strategies to enable energy-efficient operation with acceptable detection accuracy.
  • Energy expenditure estimator design: design estimators to determine energy expenditure of individuals.

This website summarizes the research efforts of Dr. Urbashi Mitra's research group (Dr. Marco Levorato, Dr. Gautam Thatte, Dr. Urbashi Mitra, Daphney-Stavroula Zois). Here, you will find our publications on the problem of energy efficient sensor selection, which is an instance of the active hypothesis testing problem for time-varying processes, and the problems related to the other three research thrusts.


Figure 1: KNOWME network schematic.

Talks

  • D.-S. Zois, ‘‘Active state tracking in heterogeneous sensor networks via controlled sensing,’’ presented at Cognitive and Algorithmic Decision-Making Seminar, University of Illinois, Urbana-Champaign, 30 April, 2014. (invited)

  • D.-S. Zois, ‘‘Active state tracking in heterogeneous sensor networks,’’ presented at the University of British Columbia, Vancouver, Canada, April 17, 2014. (invited)

  • D.-S. Zois, ‘‘Active state tracking in heterogeneous sensor networks via controlled sensing,’’ presented at USC EE CommNetS Seminar, Los Angeles, CA, February 19, 2014. (invited)

  • D.-S. Zois, ‘‘A Unified Framework for Energy Efficient Physical Activity Tracking,’’ presented at Asilomar Conference on Signals, Systems, and Computers, Pacific Grove, CA, November 4, 2013. (invited)

  • D.-S. Zois, ‘‘Non-linear smoothers for discrete-time, finite-state Markov chains,’’ presented at ISIT, Istanbul, Turkey, July 11, 2013.

  • U. Mitra, ‘‘Unified herding of CaTs: Control and Tracking in POMDPs with Gaussian Observations,’’ presented at 2013 Information Theory and Applications (ITA) Workshop, San Diego, CA, February 10-15, 2013. (invited)

  • U. Mitra, ‘‘Recursive Kalman-type state estimation within a POMDP framework,’’ presented at Cognitive and Algorithmic Decision-Making Workshop, Monticello, IL, October 5-6, 2012. (invited)

  • D.-S. Zois, ‘‘Heterogeneous Time-Resource Allocation in Wireless Body Area Networks for Green, Maximum Likelihood Activity Detection,’’ presented at ICC, Ottawa, Canada, June 12, 2012.

  • U. Mitra, ‘‘Telehealth: Opportunities for (Communications) Research?,’’ presented at 2012 Communication Theory Workshop, Maui, HI, May 14-16, 2012. pdf

  • D.-S. Zois, ‘‘A POMDP Framework for Heterogeneous Sensor Selection in Wireless Body Area Networks,’’ presented at INFOCOM, Orland, FL, March 26, 2012.

  • U. Mitra, ‘‘Next-Generation Wireless Communications: Telehealth,’’ presented at 2012 NAE National Meeting, Irvine, CA, February 2012.

  • D.-S. Zois, ‘‘POMDP Framework for Optimal Sensor Selection and Activity Detection in Wireless Body Area Networks,’’ presented at IWSM, Stanford, CA, June 11, 2011. (invited)

  • U. Mitra, ‘‘KNOWME: a wireless body area sensing network for biometric signal analysis,’’ presented at Philips, Eindhoven, March 11, 2010.