In various applications of wireless sensor networks, a high level of accuracy is required in time synchronization to maintain time consistency of sensing data as well as to reduce the idle listening times. In addition, to cope with node failures, robustness to changes in network topologies is important. In this paper, we propose two fully distributed protocols based on algorithms of multi-agent consensus. They employ an event-based approach studied in the area of networked control to enhance communication efficiency. We analyze their convergence properties and show that these protocols can achieve clock synchronization with less communication at a precision determined by the communication scheme.
Habib AissaouaMakhlouf AliouatAhcène BounceurReinhardt Euler