Improving Monte Carlo localization algorithm using time series forecasting method and dynamic sampling in mobile WSNs

Abstract

Localization of sensor nodes is one of the important operations in wireless sensor networks. Because the data produced by sensor nodes should also provide geographical location of these nodes. So having a reliable localization algorithm is always necessary. Most of presented algorithms for localization of sensor networks considered situations that the sensor nodes are static. In some of sensor networks, the nodes are mobile. So, using static localization algorithms in these networks is not suitable. Thus to support the mobility of nodes in these networks a localization algorithm will be needed that must be consistent with the mobility of nodes. Two important localization algorithms that presented in this area are Monte Carlo localization algorithm (MCL) and its improvement Monte Carlo localization boxed (MCB). Despite having a good localization accurately, sampling in these algorithms is static and they have high energy consumption. Also these algorithms are not able to localize sensor nodes in some circumstances. The main reason is that in some time slots the node can not hear any seed node. In this paper a new method has been suggested that uses forecasting and dynamic sampling for localization. This method has the ability of nodes localization in these conditions and that is an energy efficient method. Simulation results showed that the proposed method has a better performance in sparse networks in comparison with previous similar methods.