|
|
<<11/6(一)電機/資工系 聯合專題演講資訊如下>>
講題： Optimal Clustered Architectures for the Collection of Data in Large-Scale IoT Applications
演講者：Prof Edward Coyle, IEEE Fellow School of ECE, Georgia Institute of Technology, USA
時間：106年11月6日（一）14:00~16:00
地點：理工二館-第一講堂A127
Title: Optimal Clustered Architectures for the Collection of Data in Large-Scale IoT Applications
Abstract: In large-scale Internet of Things (IoT) applications, making good decisions requires both fast and energy-efficient approaches to gathering data. This is a significant challenge when the data that must be gathered is spread across thousands of devices. To demonstrate this, we consider algorithms for gathering data from the smartphones of thousands of people at large events, such as a football game. When aggregation of data is possible, clustered communication architectures for gathering that data are appropriate. We thus derive multi-level clustering algorithms that minimize the energy expended when different levels of aggregation are required at different levels of the architecture. We also show how data can be gathered in a time-efficient manner in a single cluster by only collecting the most reliable data. The next step in this research must be to determine the most energy efficient approach to data collection when there is a fixed time to gather data from all clusters, not just one cluster.
Acknowledgement: This is joint work with two of my PhD students, Deepa Phanish and Seksan Laitrakun, and with the Stadium-IoT VIP team at Georgia Tech.
|
