Low-cost wireless sensor network time synchronisation with energy conservation

An efficient time synchronisation could increase the performance of wireless sensor networks (WSNs) applications significantly. Most existing research of time synchronisation in WSNs focuses on clock estimation and message exchange to improve the synchronisation accuracy without taking network performance as the primary consideration. In this thesis, we will investigate how to extend the network lifetime while maintaining a high level of network performance via energy conservation and time synchronisation mechanisms.

To balance energy consumption and clock time accuracy in WSNs, this thesis proposes a new approach of time synchronisation that takes into account of the energy conservation factors, which reduces the energy consumption of individual nodes and extends the lifetime of WSNs by aligning schedules and reducing management packets (Chapter 3). As a cross-layer mechanism, the implementation of the proposed time synchronisation is carried out by the MAC (Medium Access Control) and application layers.

To improve the network energy conservation, a new wireless sensor network MAC protocol, named SA-MAC (Schedule Awareness MAC), is proposed. It better supports the scheduling mechanism with dimensionality reduction to generate concentrated schedule tables with active time slots information of the two-hop neighbours. The energy conservation of sensor nodes is based on the theory of periodic activation and sleep, which relies on schedule tables with accurate timestamps. The schedule tables are maintained by frequent time synchronisations (Chapter 4).

To maintain the clock accuracy of individual wireless sensor nodes, a clock adjustment method (TimeSync) for WSNs is proposed, which includes the system model, the clock offset estimation and the clock skew estimation (Chapter 5).

To balance the network energy and extend the network lifetime, a simplified reverse firefly method (SRF) was proposed. The process of the method relieves both the timeslot conflict problem and the node movement problem with energy level information exchange and comparison (Chapter 6).

By designing and implementing a set of simulation experiments, we show that by integrating the three methods above (SA-MAC, TimeSync and SRF), the energy consumption of the WSNs is reduced by more than 10% compared to the state-of-the-art MAC protocols 802.11 and S-MAC with TPSN. In the meantime, the clock accuracy of the individual nodes is maintained, which is similar to the clock accuracy achieved by TPSN.

In conclusion, the benefits of the proposed methods on WSN energy conservation and time synchronisation are discussed and the future works of clock offset estimation, node lifetime extension and network maintenance are highlighted.