Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level



Electrical Engineering


Jingxian Wu

Committee Member

Roy McCann

Second Committee Member

Yue Zhao

Third Committee Member

Qinghua Li


Distributed Estimation, Gaussian Process, Greedy Algorithms, Level Set Estimation, Sparse Sensing, Wireless Sensor Networks


This dissertation focuses on the development of theories and practices of energy aware sparse sensing schemes of random fields that are correlated in the space and/or time domains. The objective of sparse sensing is to reduce the number of sensing samples in the space and/or time domains, thus reduce the energy consumption and complexity of the sensing system. Both centralized and decentralized sensing schemes are considered in this dissertation.

Firstly we study the problem of energy efficient Level set estimation (LSE) of random fields correlated in time and/or space under a total power constraint. We consider uniform sampling schemes of a sensing system with a single sensor and a linear sensor network with sensors distributed uniformly on a line where sensors employ a fixed sampling rate to minimize the LSE error probability in the long term. The exact analytical cost functions and their respective upper bounds of these sampling schemes are developed by using an optimum thresholding-based LSE algorithm. The design parameters of these sampling schemes are optimized by minimizing their

respective cost functions. With the analytical results, we can identify the optimum sampling period and/or node distance that can minimize the LSE error probability.

Secondly we propose active sparse sensing schemes with LSE of a spatial-temporally correlated random field by using a limited number of spatially distributed sensors. In these schemes a central controller is designed to dynamically select a limited number of sensing locations according to the information revealed from past measurements,and the objective is to minimize the expected level set estimation error.The expected estimation error probability is explicitly expressed as a function of the selected sensing locations, and the results are used to formulate the optimal sensing location selection problem as a combinatorial problem. Two low complexity greedy algorithms are developed by using analytical upper bounds of the expected estimation error probability.

Lastly we study the distributed estimations of a spatially correlated random field with decentralized wireless sensor networks (WSNs).

We propose a distributed iterative estimation algorithm that defines the procedures for both information propagation and local estimation in each iteration. The key parameters of the algorithm, including an edge weight matrix and a sample weight matrix, are designed by following the asymptotically optimum criteria. It is shown that the asymptotically optimum performance can be achieved by distributively projecting the measurement samples into a subspace related to the covariance matrices of data and noise samples.