Scalable, Ad Hoc Deployable, RF-Based Localization

Abstract

Spatial localization or the ability to locate nodes is an important building block for next generation pervasive computing systems, but a formidable challenge, particularly, for very small hardware and energy constrained devices, for noisy, unpredictable environments and for very large ad hoc deployed and networked systems. In this paper, we describe, validate and evaluate in real environments a very simple self localization methodology for RF-based devices based only on RF-connectivity constraints to a set of beacons (known nodes), applicable outdoors. Beacon placement has a significant impact on the localization quality in these systems. To self-configure and adapt the localization in noisy environments with unpredictable radio propagation vagaries, we introduce the novel concept of adaptive beacon placement. We propose several novel and density adaptive algorithms for beacon placement and demonstrate their effectiveness through evaluations. We also outline an approach in which beacons leverage a software controllable variable transmit power capability to further improve localization granularity. These combined features allow a localization system that is scalable and ad hoc deployable, long-lived and robust to noisy environments. The unique aspect of our localization approach is our emphasis on adaptive self-configuration.