Research Interests

 

         Data fusion, wireless sensor networks

         Distributed Compressive Sensing

         Communications: spectrum issues, cognitive radio networks

         Network security and Byzantines, Game theory

         Medical imaging

         Stochastic resonance, Alzheimer disease detection

         Human decision making (Psychology and cognitive science)

         Acoustic sensing for ecology studies

 

  Current Research Projects

 

         Distributed Compressive Sensing, NSF

o   Development of efficient algorithms for statistical inference with low dimensional signal models where complete signal recovery is not necessary

o   Exploitation of joint sparse and structured properties of multiple data vectors to develop distributed, decentralized solutions by using low dimensional measurements at different sites, without the need to converge all the signals at a central location

o   Exploration of artificial noise injection techniques for enhancing secrecy performance in the presence of an eavesdropper

 

         Heterogeneous Multimodal Data Fusion, ARO

o   Heterogeneous data from multimodal sensors is statistically modeled using copula theory literature

o   Data fusion methods for statistically dependent data/features/decisions are explored for detection and estimation

o   Information diversity in heterogeneous sensor networks is characterized and analyzed

 

         Resource Management in Sensor Networks, AFOSR

o   Sensor scheduling problem with the goal to reduce the number of selected sensors, is explored using sparsity-aware optimization

o   Equilibrium-based two-sided market model is formulated to design a crowdsourcing based mechanism for the distributed inference using sensors that are selfish and profit-motivated

o   Probabilistic sensor management schemes are explored based on the concepts developed in compressive sensing

 

         Network Security, NYSTAR, AFRL

o   Design of schemes for interference management in wireless communication networks via strategic power allocation, matching/market theory, and coding-based approaches

o   Analysis of optimal Byzantine attacks and network's defense strategies in distributed inference networks with different topologies

o   Design of secure distributed detection systems in the presence of an eavesdropper

o   Development of a tool to facilitate analysis of strategic use of redundancy and diversity techniques for cyber survivability and recoverability against tactical attacks by leveraging Game Theoretic concepts

 

         Interdisciplinary Collaborations, NSF

o   Design of automatic large-scale classifiers for bird sounds using large data that has been collected by ecologists to study bird behavior patterns

o   Development of experiments in collaboration with psychologists to understand the phenomenon of decision fusion by humans

 

Last Updated: February, 2015