1: \begin{abstract}

2: Recently, the distributed state estimation problem for continuous-time linear systems over jointly connected switching networks was solved.

3: It was shown that the estimation errors will asymptotically converge to the origin by using the generalized Barbalat's Lemma.

4: This paper further studies the same problem with two new features.

5: First, the asymptotic convergence is strengthened to the exponential convergence.

6: This strengthened result not only offers a guaranteed convergence rate,

7: but also renders the error system total stability and thus is able to withstand small disturbances.

8: Second, the coupling gains of our local observers can be distinct and thus offers greater design flexibility,

9: while the coupling gains in the existing result were required to be identical.

10: These two new features are achieved by establishing exponential stability for two classes of linear time-varying systems,

11: which may have other applications.

12: 

13: \end{abstract}

14: