Fine Synchronization for Wireless Sensor Networks Using Gossip Averaging Algorithms

Abstract

International audienceAs synchronization preambles represent a great overhead on wireless data packets, it could be interesting to powerful them. Fine synchronization solves this problem by homogenizing clock drift values on network regions. When coupled with any clock bias consensus method (coarse synchronization), fine synchronization allows the setup of time-based medium multiplexing schemes (TDMA, Slotted-ALOHA, ...) . This paper deals with a lightweight method able to perform the computation of local clock drift corrections using simple interaction rules between neighbor nodes. The solution proposed in this work has the main advantage to be robust to measurement noise while having the technologically interesting ability to drive clock drifts towards an average consensus

Joint Time Synchronization and Localization of an Unknown Node in Wireless Sensor Networks

“…Solving (21) in LS sense, we can obtain an estimate of as (22) However, note that (22) only provides a rough estimate of the unknown vectors and , because of two reasons. First, the estimator (22) assumes all the components in the measurement error vector have the same variances, while it is obviously not the case as can be seen from (17). Second, there is no constraints applied between the elements of in (22), and thus the estimate may be inconsistent.…”

“…Therefore, from (17), can be approximated by , where . The covariance matrix of the noise vector can then be derived as .…”

“…Notice that the hierarchical time synchronization and localization method we consider in this paper is a centralized approach in the sense that the procedure starts from the anchors and proceeds to the next level. It is different from many distributed time synchronization techniques, such as clock consensus and gossip averaging, where no reference node is needed [17]- [19].…”

A New Distributed Approach for Achieving Clock Synchronization in Heterogeneous Networks

“…Many different schemes have been proposed that seek to achieve timing synchronization in two-tier networks [1]. These methods can be divided into approaches that take advantage of synchronization sources within the network (internal sensing), e.g., the signal from neighbouring macrocells [1], femtocells [8], [9], and mobile users [10]- [12] and also the backbone connection [13]- [15] or algorithms that employ synchronization sources from outside the network (external sensing), e.g., global positioning systems (GPS) [1] and TV signals [16]. The pros and cons of each scheme are briefly summarized below:…”

“…• In [8], [9], the signal from neighbouring femtocells is used to achieve synchronization throughout the two-tier network. However, such an approach is only possible if the areas of coverage of many femtocells overlap one another.…”