Bound analysis for anchor selection in cooperative localization

Abstract

Anchor selection refers to choosing a small portion of the nodes with known locations to ensure the unique localizability and/or improve the accuracy of cooperative localization. Focusing on the localization accuracy, conventional practice suggests that the anchors should be deployed on the perimeter of the network. This paper derives the perimeter anchor deployment strategy by performing a bound analysis for the Cramér-Rao lower bound (CRLB) which quantifies the localization accuracy. It is proved that the uniform perimeter anchor deployment strategy is the optimal to fix an isotropically discriminable relative configuration whose nodes are randomly deployed onto a two-dimensional plane. For the relative configuration specified by the internode distance measurements, we introduce an error metric to evaluate the anchor selection performance, together with an upper bound that is independent of anchor selection.