Issues in Underwater Acoustic Signal Processing in the Era of Big Data
As many unmanned sensors/platforms are likely to be deployed in the ocean in the future to collect data, we are entering an era of big data for underwater sensing.The associated problems, the corresponding issues on data/signal processing to extract the desired information for underwater data are more problematic compared with the big data presented in other disciplines such as web data management or cloud computing.For big data in business, banking, web management, cloud computing,etc., the data can be considered as structured, namely, each entity has a finite set of parameters that are pre-defined to describe them.On the other hand, the data collected underwater can be associated with in principle an infinite number of possible descriptions although in practice, the possibilities can be limited to a finite but still very large set.As a result, more may not be better or helpful without putting some smart into the system.How to select the ”smart” as a cue for data/signal processing becomes an important issue in order to reduce the data to small useful finite sets for signal processing and information extraction.In analogy to target detection, how to detect the relevant information bearing data (for the desired information) is a research issue that awaits to be explored.Without that, brute force signal processing for each incoming data set is not realistically possible due to the computations and energy load required.
Big data underwater sensing data/signal processing
YANG T.C.
College of Information Science and Electronic Engineering,Zhejiang University, Hangzhou 310027
国内会议
杭州
英文
2-13
2015-12-01（万方平台首次上网日期，不代表论文的发表时间）