Date of Award
Master of Science in Electrical Engineering
Department of Electrical and Computer Engineering
Richard K. Martin, PhD
Radio tomographic imaging (RTI) is a form of device-free, passive localization (DFPL) that uses a wireless sensor network (WSN) typically made up of affordable, low-power transceivers. The intent for RTI is to have the ability to monitor a given area, localizing and tracking obstructions within. The specific advantages rendered by RTI include the ability to provide imaging, localization, and tracking where other well developed methods like optical surveillance fall short. RTI can function through optical obstructions such as smoke and even physical obstructions like walls. This provides a tool that is particularly valuable for tactical operations like emergency response and military operations in urban terrain (MOUT). Many methods to optimize the performance of RTI systems have been explored, but little work that focuses on the sequence of transceiver reports can be found in the body of literature. This thesis provides an exploration of the effects from attempting to optimize the transmission sequence in a WSN by creating a metric to quantify the value of the information a transceiver will report and using it to develop a dynamic, utility-driven, token passing process. After deriving a metric from the Fisher information matrix of the imaging solution, it was combined with a weighting based on the time each node last reported across the WSN. Modeling and simulation was performed to determine if the novel transmission sequence provided any benefit to the localization and tracking performance. The results showed a small improvement in two different localization methods when packet loss in the WSN reached 50%. These results provide a proof-of-concept that warrants further exploration and suggest that performance improvements may be realized by implementing a transmission sequence based on the metric developed in this thesis.
Nafziger, Grant T., "Wireless Sensor Network Optimization for Radio Tomographic Imaging" (2020). Theses and Dissertations. 3185.