1: \begin{abstract}

2: 

3: Contraction analysis is a stability theory for nonlinear systems where

4: stability is defined incrementally between two arbitrary

5: trajectories. It provides an alternative framework in which to study

6: uncertain interconnections or systems with external inputs, where it

7: offers several significant advantages when compared with traditional

8: Lyapunov analysis. Contraction-based methods are particularly useful

9: for analyzing systems with uncertain parameters and for proving

10: synchronization properties of nonlinear oscillators. Existence of a

11: contraction metric for a given system is a necessary and sufficient

12: condition for global exponential convergence of system

13: trajectories. For systems with polynomial or rational dynamics, the

14: search for contraction metrics can be made fully algorithmic through

15: the use of convex optimization and sum of squares (SOS)

16: programming. The search process is made computationally tractable by

17: relaxing matrix definiteness constraints, whose feasibility indicate

18: existence of a contraction metric, into SOS constraints on polynomial

19: matrices.  We illustrate the results through examples from the

20: literature, emphasizing the advantages and contrasting the differences

21: between the contraction approach and traditional Lyapunov techniques.

22: 

23: \end{abstract}

24: