专用集成电路与系统国家重点实验室
讲座信息

报告人：Prof. Sheng-Guo Wang（University of North Carolina at Charlotte）
时    间：2016年5月18日下午2:00-4:00
地    点：张江校区微电子楼389室

Seminar Topic 1:
“Wetland Identification and Its Automation based on DEM Derivatives and Machine Learning Random Forest Method"

(That is his last year conference paper "Random Forest Classification and Automation for Wetland Identification based on DEM Derivatives" at 2015 ICOET, Sept. 2015.）

Abstract:
This speech presents his recently published research work: wetland identification and its automation process based on DEM derivatives and random forest machine learning method. It not only dramatically improves the initial modeling efforts based on DEM derivatives by random forest (RF) classification method, and but also provides a full automation process to wetland identification, that exemplify how innovative technologies can be used in lieu of extensive field wetland delineations and ultimately reduce transportation project delivery time and costs while protecting the environment.  It deals with big data and utilizes machine learning RF method for modeling and prediction in wetlands identification.
The results show that the RF method significantly improves the prediction accuracy for wetlands by reducing error about 10 times comparing with the Logistic regression method, and the automation process as an innovation makes the whole prediction process vividly easy and impressive speed. 

Seminar Topic 2:
“Novel Lyapunov-type Functional for Robust Control of General Time-Varying Delay Stochastic Uncertain Systems with Brownian Motion”

 
Abstract:
This talk presents a novel Lyapunov-type functional for robust control to complex systems with Brownian motion, i.e., Wiener process. The considered complex systems may include time-varying delay, structured and unstructured uncertainties, and Wiener process as a time-varying delay stochastic uncertain systems.
The novel Lyapunov-type functional is not a Riemann integral, but a Stieltjes-type double integral that is an essential difference from all previous types of Lyapunov functional, making it novel. It is much general and able to work on general systems, but previous ones may not.
Thus, the presented new Lyapunov-type functional and its combination with the common Lyapunov functional may have broad applications in view of its validity to broad complex systems, including its simplified systems.

Bio:
The speaker Dr. Sheng-Guo Wang is a professor at University of North Carolina at Charlotte. He received his B.S. and M.S. in electrical engineering from University of Science and Technology of China in 1967 and 1981 respectively, and PhD in electrical and computer engineering from University of Houston in 1994.
He has been the PI for numerous research projects since 1974. Prof. Wang is a recipient of China National Science Conference Prize 1978 (one of the highest academic awards in China) and many other academic awards, e.g., his recent NCDOT research project, titled “Improvements to NCDOT’s Wetland Prediction Model” (4-30-2012 ~ 8-15-2014), has won a national “2015 Sweet 16 High Value Research” award recognized by AASHTO (American Association of State Highway and Transportation Officials) and RAC (Research Advisory Committee), 2015, and also been acknowledged at 2016 TRB (Transportation Research Board) Annual Meeting, National Academies of Sciences-Engineering-Medicine, Jan. 11, 2016.

 
联系人：曾璇、严昌浩