The primary idea behind deploying sensor networks is to utilize the distributed sensing capability provided by tiny, low powered and low cost devices. Multiple sensing devices can be used cooperatively and collaboratively to capture events or monitor space more effectively than a single sensing device. The realm of applications envisioned for sensor networks is diverse including military, aerospace, industrial, commercial, environmental and health monitoring. Typical examples include: traffic monitoring of vehicles, crossborder infiltration-detection and assessment, military reconnaissance and surveillance, target tracking, habitat monitoring and structure monitoring, to name a few. Most of the applications envisioned with sensor networks demand highly reliable, accurate and fault-tolerant data acquisition process. The integrity of data alone can have tremendous effects on the performance of any data acquisition system. Due to the low manufacturing cost, the sensors lend themselves to be deployed in large numbers with a high spatial distribution. Such a large deployment scheme often generates enormous amount of data that needs to be efficiently summarized and delivered for analysis and processing. In-network data compression, data aggregation/fusion, and decision propagation are some of the processes that deal with huge data issues. A hierarchical data aggregation scheme developed in this thesis is a highly effective and energy efficient means (by reducing communication packets) to deliver decision milestones to the enduser. The sensing devices are also prone to failure due to the inherent characteristics such as construction and deployment. It is thus necessary to devise a fault-tolerant mechanism with a low computation overhead to validate the integrity of the data obtained from the sensors. Moreover, a robust diagnostics and decision making process should aid in monitoring and control of critical parameters to efficiently manage the operational behavior of a deployed sensor network. Specifically, this research will focus on innovative approaches to deal with multi-variable multi-space problem domains (data integrity, energy-efficiency and fault-tolerant framework) in wireless sensor networks. We present three information-based methods for improving the performance (faulttolerance and efficiency) of wireless sensor networks (WSNs). The first is a method for time varying weight adaptation in a mixture model for sensor data aggregation. The second technique applies fuzzy inference methods to solve a multi-criteria decision problem, specifically the efficient management of data collection in a WSN. The third method presented proposes the use of spatially variant weights to reduce the significance of sensor readings taken near the boundary of the sensor range, in order to minimize potential corruption of aggregated data. The solutions proposed in this thesis have practical implementation in developing power-aware software components for designing robust networks of sensing devices.
Level of Degree
Electrical and Computer Engineering
First Committee Member (Chair)
Second Committee Member
Third Committee Member
Fourth Committee Member
Sridhar, Prasanna. "Hierarchical aggregation and intelligent monitoring and control in fault-tolerant wireless sensor networks." (2007). https://digitalrepository.unm.edu/ece_etds/243