Research On The Localization Of Mobile Nodes In Wireless Sensor Networks Using Monte Carlo Method

Wireless sensor networks, which integrate the three technologies of sensor, micro-electronics and network, is a whole new technology of obtaining and processing information. And it has been applied in the fields of military, disaster overcoming, environment monitoring, medical care and space discovering. The data is meaningless for most application without knowing the sensor's position. Due to the cost, form size, power and circumstances, it is unrealistic s to equip all nodes with GPS localization system. Therefore, it is significant to research the self-localization of wireless sensor nodes.The localization technology of nodes can classified as Range-based and Range-free Localization according to localization methods. Range-based localization employs the equipped hardware to measure the angle or distance between nods, and use the triangulation, the trilateration or the multi-lateration method to compute node's position. Range-free localization then make use of the principles of proximity or hops in distance to realize node's localization.Most of the current localization methods, which have not considered the node's mobility at the design stage, suffer from mobility. But Monte Carlo Localization is aimed at dealing with node's mobility, and its essential idea is to employ N samples to denote the posterior probability distribution. It can make better use of the node's speed, direction and connectivity than other methods, but it also has two shortcoming: 1) the sampling region is large and leads to big localization error; 2)the estimation of unknown nodes has not been used. This paper proposed MCBE method to overcome the two shortcomings. MCBE use the position of one-hop and two-hop beacon nodes to get an anchor box, and the sampling is confined to the box to reduce the sample area. In addition, the size of anchor box is close related to the localization error. The position estimations of nodes which have small anchor boxes are used to reduce the neighbors'anchor boxes that are larger than some certain value. In the end, MCBE is simulated and is contrasted with MCL and MCB in the aspects of localization error, localization coverage and communication cost. MCBE has the least localization error, highest localization coverage and greatest communication cost.