Methodological remarks on contraction theory

被引：64

作者：

Jouffroy, J ^{[1
]}

Slotine, JJE ^{[1
]}

机构：

[1] NTNU, CESOS, NO-7491 Trondheim, Norway

来源：

2004 43RD IEEE CONFERENCE ON DECISION AND CONTROL (CDC), VOLS 1-5 | 2004年

关键词：

D O I：

10.1109/CDC.2004.1428824

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Because contraction analysis stems from a differential and incremental framework, the nature and methodology of contraction-based proofs are significantly different from those of their Lyapunov-based counterparts. This paper specifically studies this issue, and illustrates it by revisiting some classical examples traditionatly addressed using Lyapunov theory. Even in these cases, contraction tools can often yield significantly simplified analysis. The examples include adaptive control, robotics, and a proof of convergence of the deterministic Extended Kalman Filter.

引用

页码：2537 / 2543

页数：7

共 20 条

[1] Natural gradient works efficiently in learning [J].