Recursive position estimation in sensor networks 论文

摘要

Recursive hierarchy provides a framework for extending position estimation throughout a sensor network. Given imprecise ranging and inter-node communication, nodes scattered throughout a large volume can estimate their physical locations from a small set of reference nodes using only local information. System coverage increases iteratively, as nodes with newly estimated positions join the reference set, capitalizing on the massive scale of sensor networks. The system frames position estimation as a geometric problem solvable through common nonlinear regression techniques and develops methods for gauging the reliability of position estimates. This provides a flexible framework that can use and enhance a variety of technologies and protocols to produce fine-grained position estimates. A specific model provides a simulation environment showing that over 90% of position estimates are correct to within 3% of the ranging distance with only 5% of the system in the initial reference set.