金沙娱城总站3777|首页-welcome!

         
 
   
-->
首页 --> 公告通知

Robust Analysis of Consensus Algorithms

创建时间:  2019-12-31  王智渊    浏览次数:


报告时间:2020年1月12日 (周日)10:00-11:00   

报告地点:宝山校区机自大楼702B


报告人: 纪成达, 博士, 约翰霍普金斯大学

邀请人: 任肖强 教授


Title: Robust Analysis of Consensus Algorithms


Abstract: A fundamental problem in the networked dynamical system is to achieve the consensus in the presence of disturbance. In this talk, we present our recent progress towards developing distributed control strategies for the design of consensus systems. Our approach is to analytically evaluate the robust performance through the input-output norm of the dynamical system. A parametrized family of consensus algorithms is studied. In particular, we show that the input-output norm is finite unless the zero eigenvalue of the system is not observable from the output. Two special consensus applications, vehicle collision potential and sensor failure contingency analysis, are presented to illustrate our approach. We further propose two easy-to-implement approaches to improve consensus performance. In the first approach, which is referred to as the augmented consensus algorithm, a consensus filter is introduced in the feedback loop that transfers the consensus-oriented estimation to the system. In the second approach, which is referred to as the robust consensus algorithm, we add time-relative feedback from the current state to the initial state.


Bio: Chengda Ji graduated from a joint program and received his Bachelor of Engineering and Bachelor of Science degrees in Mechanical Engineering from Beijing Institute of Technology, China, and Polytechnic University of Turin, Italy, respectively, in 2016. He then joined the Department of Mechanical Engineering at Johns Hopkins University as a research assistant with Professor Dennice Gayme as his Ph.D. advisor. His research interests include modeling, dynamics, and control of networked systems, e.g., power systems, vehicle platoons, and computer networks.  His current research focuses on developing distributed control and learning frameworks for networked systems. Chengda was awarded the Johns Hopkins Graduate Fellowship (2017) and CSC Undergraduate Fellowship (2015).






上一条:Dynamic Droop Control in Low-inertia Power Systems
下一条:111引智基地讲坛第25讲-Topic: Applications & Future Trends of Advanced Industrial Robots



版权所有 ©       地址:上海市宝山区上大路99号   邮编:200444    
技术支持:
联系我们