学术报告
同济大学计算机科学与技术系智信讲坛第（103）期
题目：Non-intrusive appliance load monitoring using graph signal processing
报告人：Vladimir Stankovic
时间：2017年7月5日星期三下午4:00--5:00
地点：电信楼215
组织单位：计算机科学与技术系
邀请人：郑思明
报告人简介：
Dr Vladimir Stankovic is a Reader in the Dept. of Electronic and Electrical Engineering, Faculty of Engineering, at the University of Strathclyde, Glasgow in Scotland, UK with over 15 years of research experience in digital signal processing and communications. He has co-authored 4 book chapters, 60 peer-reviewed journal papers, over 100 papers in peer-reviewed intl. conferences (h-index 29, google scholar), and has been awarded 5 patents. He has led several projects funded by European Union, Research Council UK, and Scottish Funding Council. He is Editor-At-Large of IEEE Transactions on Communications, Associate Editor of IEEE Transactions on Image Processing, Area Editor of Elsevier Signal Processing: Image Communications Journal, and General Chair of IEEE Multimedia Signal Processing (MMSP) Workshop 2017.
内容提要：
Energy disaggregation via non-intrusive appliance load monitoring (NILM) refers to breaking total household/building energy consumption down to individual appliances used. NILM has many applications including enhancing energy feedback, supporting smart home/building automation, assisted living, and so on. This talk will start by giving a short overview of NILM methods from its origin in early 1980’s when NILM was introduced as a communication over noisy channel problem, until nowadays, highlighting major challenges and remaining research opportunities. The talk will mainly focus on the most challenging NILM problem based on conventional low-rate smart meters that are being deployed at large scale for billing purposes.
In the second part of the talk, graph signal processing (GSP) will be introduced and its potential for low-rate NILM will be demonstrated.