Relative localization with 2-hop neighborhood

Abstract

Localization is the process in which nodes in a wireless sensor network self-determine their positions in the network. While there are many effective mathematical techniques for solving the problem of localization, most are not suitable for the resource-constrained distributed environment of sensor networks. We propose ANIML an iterative, range-aware relative localization technique for wireless sensor networks that requires no anchor nodes. ANIML restricts itself to the use of only local 1- and 2-hop neighbor information, avoiding the need for information flooding and thus controlling cascading ranging errors that bedevil other localization techniques. While least-squares minimization is a mathematically simple constraint optimization technique, utilizing 1- and 2-hop neighbor information as constraints, ANIML provides better localization without the need for more sophisticated error control and/or global information. We implemented ANIML in ns-2 and conducted extensive experimentation to evaluate its performance. Experimental results show that ANIML provides robust localization and scales well.