User Tools

Site Tools

Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/3/3121c14bad4a7110989ed6cd98048df8.i failed
Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/3/3121c14bad4a7110989ed6cd98048df8.metadata failed
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/3/3121c14bad4a7110989ed6cd98048df8.i failed
Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/3/3121c14bad4a7110989ed6cd98048df8.xhtml failed

Title: Active state tracking in heterogeneous sensor networks via controlled sensing

Abstract: Active state tracking constitutes a generalization of the classical state tracking problem. In particular, the objective is to determine a way of managing sensing resources (e.g. sensor types, number of samples, location, etc.) to accurately and efficiently track the unknown state of a dynamical system. Thus, in contrast to traditional control systems, where the control input affects the system state evolution, in active state tracking applications, the controller actively selects between the available observations, but does not affect the plant. Active state tracking applications include but are not limited to: sensor scheduling for target detection and tracking, health care, generalized search, waveform selection for radar imaging, estimation of sparse signals and coding with feedback.

In this talk, I am going to focus on the active state tracking problem for systems modeled by discrete-time, finite-state Markov chains. The 'hidden' system state is observed through noisy Gaussian observation vectors that depend both on the underlying state and an exogenous control input, which shapes their quality. To accurately track the time-evolving system state, we address the joint problem of determining recursive formulae for minimum mean–squared error (MMSE) state estimators and designing a control strategy. To avoid the computational burden associated with the optimal control strategy, we consider alternative ways of determining the optimal control policy. Specifically, for the case of two states and scalar measurements, we derive properties of the related cost functions and characterize sufficient conditions regarding the structure of the optimal control policy.

Bio: Daphney-Stavroula Zois received the B.S. degree in computer engineering and informatics from the University of Patras, Greece, in 2007, and the M.S. degree in electrical engineering from the University of Southern California (USC), Los Angeles, in 2010. Currently, she is a Ph.D. candidate in electrical engineering at USC working with Prof. Urbashi Mitra on active state tracking problems. Ms. Zois is a recipient of several fellowships and awards including the Ming Hsieh Scholar award for 2012–2013, the Panagiotis Triantafyllidis scholarship, the Myronis Graduate fellowship, the Greek Women’s Engineering Association award, the Viterbi Dean’s fellowship, and the Andreas Mentzelopoulos scholarship. Her research interests include intelligent control, detection, estimation, and signal processing in sensor networks, and machine learning.

active_state_tracking_in_heterogeneous_sensor_networks_via_controlled_sensing.txt · Last modified: 2016/09/01 19:15 (external edit)