1: \begin{abstract}

2:   We propose a distributed algorithm for time synchronization in

3:   mobile wireless sensor networks. Each node can employ the algorithm

4:   to estimate the global time based on its local clock time. The

5:   problem of time synchronization is formulated as nodes estimating their skews and

6:   offsets from noisy difference measurements of offsets and logarithm

7:   of skews; the measurements acquired by time-stamped message

8:   exchanges between neighbors. A distributed stochastic approximation based algorithm is

9:   proposed to ensure that the estimation error is mean square

10:   convergent (variance converging to 0) under certain conditions. A sequence of scheduled update

11:   instants is used to meet the requirement of decreasing

12:   time-varying gains that need to be synchronized across nodes with

13:   unsynchronized clocks. Moreover, a modification on the  algorithm is

14:   also presented to improve the initial convergence

15:   speed. Simulations indicate that highly accurate global time

16:   estimates can be achieved with the proposed algorithm for long time

17:   durations, while the errors in competing algorithms increase over time. % This improvement is principally due to the

18: %   fact that the variance of the skew estimates obtained by the

19: %   proposed algorithm are asymptotically zero.

20: \end{abstract}

21: