Title:
Imaging Seismic Tomography in Sensor Networks

Abstract:
Existing seismic instrumentation systems do not yet have the capability to recover physical dynamics with sufficient resolution. At present, raw seismic data are typically logged at a few stations then manually retrieved months later for post processing and tomographic imaging at a central server. Thus neither real-time nor high-resolution tomography imaging are possible today which limited our understanding of geospatial dynamics. There is a consensus that sensor network will significantly advance the instrumentation for geospatial study. But traditional data collection paradigm cannot work well, because it is virtually impossible to collect real-time data from a large-scale wireless seismic network to a central server due to the sheer data amount, bandwidth and energy constraints. In our NSF CDI project, we are developing a VolcanoSRI (Volcano Seismic Realtime Imaging) system, a large-scale mesh network of low-cost seismic stations, that sense and analyze seismic signals, and compute real-time, three-dimensional fluid dynamics of a volcano conduit system (e.g., 4D volcano tomography) within the sensor network. Realizing such a VolcanoSRI system requires a transformative approach to tomography computation algorithm, collaborative signal processing, and the associated sensor network design. In this talk, we present our recent research on distributed tomography algorithms that process data and invert volcano tomography in the network, while avoiding costly data collections and centralized computations. The new algorithm distributes the computational burden to sensor nodes and performs realtime tomography inversion under the constraints of network resources.

Speaker Bio:
Dr. Wen Zhan Song is an associate professor of Computer Science and director of Sensorweb Research Laboratory at Georgia State University. His research mainly focuses on geospatial analytics, smart grid and smart environment where sensing, computing, communication and control play a critical role and need a transformative study. He has received $6 million+ research funding from NSF, NASA, USGS, and Boeing since 2005. Dr. Song is a recipient of the Outstanding Research Contribution Award (2012) in GSU Computer Science, the Chancellor Research Excellence Award (2010) in WSU Vancouver, and the NSF CAREER Award (2010). His research has been featured in MIT Technology Review, Network World, Scientific America, New Scientist, and National Geographic. During his PhD study, he was also a recipient of 2004 National Outstanding Oversea Student Scholarship, awarded by Ministry of Education of China (only 40 awarded in USA). Before that, he was a software engineer and team leader in Alcatel-Lucent Shanghai Bell. Dr. Song serves the editorial board of several premium journals including IEEE Transaction on Parallel and Distributed Systems.