| dc.description.abstract | Clock synchronization in a wireless sensor network (WSN) is essential as it provides
a consistent and a coherent time frame for all the nodes across the network. Typically,
clock synchronization is achieved by message passing using carrier sense multiple
access (CSMA) for media access. The nodes try to synchronize with each other, by
sending synchronization request messages. If many nodes try to send messages simultaneously, contention-based schemes cannot efficiently avoid collisions which results in message losses and affects the synchronization accuracy. Since the nodes in a WSN have limited energy, it is required that the energy consumed by the clock synchronization protocols is as minimum as possible. This can be achieved by reducing the duration for which the clock synchronization protocols execute. Synchronous clock synchronization
protocols in WSNs execute the clock synchronization process at each node, roughly
during the same real-time interval, called synchronization phase. The duration when
there is no synchronization activity is called the synchronization interval. Synchronization phases are divided into synchronization rounds. The energy consumed by these protocols depends on the duration of the synchronization phase and how frequently the synchronization phase is executed. Hence, to minimize the energy consumption by each
node, the duration of synchronization phase should be as small as possible. Due to different drift rates of the clocks, the synchronization phases at different nodes drift apart and special techniques are required to keep them in sync. An existing protocol, called improved weighted-average based clock synchronization (IWICS) uses a pullback technique to achieve this. If a message from (i + 1)th synchronization round is received by a node still executing the ith synchronization round, the receiving node reduces its next synchronization interval to ensure greater overlap in the synchronization rounds. The reduction in overlap is a gradual and continuous phenomenon, and so, it can be detected and dealt with continuously.
In this thesis, first, we make use of TDMA-based MAC protocols, instead of CSMA, to
deal with the problem of message losses. We discuss the challenges of using TDMA-based
MAC protocols for clock synchronization and how to overcome these challenges. Second,
The IWICS protocol calculates the virtual drift rate which we use to modify the duration of the synchronization interval so that there is more overlap between the synchronization phases of neighbouring nodes. We refer to this technique as drift rate correction. Finally, we propose a different pullback technique where the pullback detection is carried out in each of the synchronization phase as opposed to the old pullback mechanism where it would be detected only when an out-of-round synchronization message is received.
The proposed pullback technique when applied to the current synchronization interval
ensures that the synchronization phases, that follow the current synchronization interval,
are better synchronized with each other. As a result of this, we are able to reduce
the duration of synchronization phases further. The IWICS protocol with all these
modifications incorporated is termed as the TIWICS (TDMA-based IWICS) protocol.
Simulation and experimental results confirm that the TIWICS protocol performs better
in comparison to the existing protocols. | en_US |
