For the nodes that have received more than one time-stamped messages
originated from di?®erent master nodes, they use the standard deviations car-
230
Chapter 9 Time-Synchronization Challenges and Techniques
ried in the time-stamped messages as weighted ratio of their time contribution
to the new time. In essence, the nodes weight the times di?®used by the master
nodes to obtain a new time for them. This process is to provide a smooth time
variation between the nodes in the network. The smooth transition is important
for some applications such as target tracking and speed estimating.
The master nodes are autonomously elected, so the network is robust to
failures. Although some of the nodes may die, there are still other nodes in
the network that can self-determine to become master nodes. This feature
also enables the network to become server-less if necessary and to reach an
equilibrium time. In addition, the master and di?®usion leader nodes are selfdetermined
based on their own energy level. Also, the TDP is light-weight,
but it may not be as tunable as the RBS.
Li and Rus propose an asynchronous rate-based di?®usion algorithm [11],
which aims to converge the time in the sensor network to the average time
of the network. The asynchronous rate-based di?®usion algorithm is given in
Table 3. A sensor node obtains the clock readings from its neighboring nodes
and computes the average of these readings.
Pages:
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379