|
|
 |
| |
|
|
|
Distributed Network Localization for Wireless Sensor Networks |
|
|
|
PP: 1109-1115 |
|
|
|
Author(s) |
|
|
|
Seong-Woo Kim,
Jong Min Lee,
Jung-Hwa Lee,
|
|
|
|
Abstract |
|
|
| Recent advancements in wireless communication and microelectro-mechanical systems
(MEMS) have made possible the deployment of wireless sensor networks for many real world
applications. One of most challenging problems with the deployed sensor nodes is to identify their
geographic locations given estimates of the distances between them. There have been a large number
of localization algorithms, each of which makes a different geometric approximation. Among them,
Multidimensional scaling (MDS) based algorithms outperform the others and have the advantage that
they are robust for noise and sparse networks, with or without anchor nodes. Its distributed versions
compute a local map for each node at first and then merge these maps to a global map. Additional
refinement technique can improve the relative maps by forcing them to conform more closely to the
distances to nearby neighbors. In this paper, we reformulate the network localization problem as a
constrained least square problem mathematically in detail, and then develop a new refinement
algorithm which is based on two hop distance constraints and Levenberg-Marquardt optimization
technique. Our simulation results demonstrate that the proposed algorithm has good performance in
terms of both success rate and node position estimation.
|
|
|
|
|
|
|
|
