March 27, 2012, 1132 Harold Frank Hall

Recent advances in technology have enabled the use of wireless sensor networks for environmental monitoring and surveillance. Wireless sensor networks are particularly beneficial for monitoring environments that are unsuitable for human presence, such as those arising in the monitoring of permafrost, volcanos, forest fires, and battlefields. Mobile agents called data mules can be used to enhance sensor networks by visiting individual sensors to collect measurements. In this talk, the mobility of the data collector will be used as a way to mitigate energy depletion of the stationary nodes and allow for sparse deployments at the cost of additional data latency. This motivates us to seek efficient strategies for the mobile vehicle in order to alleviate this data latency.The policies to be described apply directly to an acoustic source localization problem in which the objective is to localize the source of a transient acoustic event using measurements from a group of unattended ground sensors. First we will discuss optimal sensor placement of acoustic time of arrival sensors for localization using techniques from information theory and optimization. A sparse sensor configuration is shown to be optimal when maximizing the expected determinant of the Fisher information matrix for truncated, radially-symmetric source distributions. Next, we will use ideas from optimal sensor placement and optimal sensor selection to adaptively adjust the route of the data mule to minimize the time to localize events of interest. When the data mule is a small fixed wing unmanned aerial vehicle (UAV), this algorithm can be improved by including the kinematic constraints of the UAV and exploring the ability of the UAV to communicate with faraway sensors within its line of sight. Finally, we will describe data mule routing policies for acoustic source localization over a large area involving many sources and sensors and analyze their performance.

Jason T. Isaacs received his bachelor’s degree in Electrical Engineering from the University of Kentucky in 1999. Upon graduation he spent the next six years working as a motion control development engineer for Lexmark International Incorporated focusing on the paper feeding systems of inkjet printers. He earned an M.S. degree in Electrical and Computer Engineering in 2008 from the University of California Santa Barbara.  Jason is currently pursuing a Ph. D. degree in Electrical and Computer Engineering from the University of California Santa Barbara under the supervision of Professor João Hespanha.  His research interests include sensor networks, estimation, and UAV routing.