Message Passing Based Time Synchronization in Wireless Sensor Networks: A Survey

Abstract

Various protocols have been proposed in the area of wireless sensor networks in order to achieve network-wide time synchronization. A large number of proposed protocols in the literature employ a message passing mechanism to make sensor node clocks tick in unison. In this paper, we first classify Message Passing based Time Synchronization (MPTS) protocols and then analyze them based on different metrics. The classification is based on the following three criteria: structure formation of the network affected by the synchronization protocol, frequency of synchronization process (synchronization interval), and synchronization message overhead. Proposed protocols are analyzed and evaluated from different perspectives based on available data. A comparison table of the reviewed protocols is presented according to the evaluation metrics. Finally, some potential methods will be proposed to improve the synchronization process.

EERS: Energy-Efficient Reference Node Selection Algorithm for Synchronization in Industrial Wireless Sensor Networks

“…It consists of a pair selection methods and hierarchy forming. GPA, however, needs an additional pairwise operation that reduces the energy efficiency compared with its single-hop counterpart [ 7 ]. Selecting pairs of nodes in a large network can be very costly in terms of both computation time and energy.…”

“…Generally, sensor nodes are frequently prepared with hardware clock oscillator, which is relatively inexpensive, and inaccurate. In this way, it is much challenging to introduce precise synchronization for such sensor nodes [ 6 , 7 ]. With such sensor nodes, it is required to periodically conduct an accurate synchronization process to well-preserved time synchronization of networks for a long time [ 8 ].…”

“…All sensor nodes, therefore, require regularly exchange timing messages in order to keep the whole network synchronized. Exchanging timing messages minimizes the time offsets that produced by the clock drift of each node in the network [ 7 , 10 ]. Frequent synchronization, however, dramatically raises energy consumption as the transmission of messages typically corresponds to the most significant portion of energy consumption [ 10 ].…”

Performance Analysis of Time Synchronization Protocols in Wireless Sensor Networks

“…Several studies [26,27,28,29] have provided comparisons of time synchronization protocols based on theoretical analyses. However, they failed to provide quantitative performance analyses due to the absence of a simulation study.…”

“…In this study, we conduct a performance analysis on the Flooding Time Synchronization Protocol (FTSP) [1] and Gradient Time Synchronization Protocol (GTSP) [11] which are representative time synchronization protocols with centralized and distributed mechanisms, respectively. Both protocols are listed in many surveys and comparative studies [25,26,27,28,29], and in particular, they are widely used as benchmarks for newly proposed time synchronization protocols [4,30]. Many protocols use similar methods to those of FTSP (least-squares regression) and GTSP (average consensus) in estimating clock drift.…”