Tradeoff between estimation performance and sensor usage in distributed localisation problems

In distributed sensor networks, the tradeoff between application performance and resource consumption is a fundamental issue. In this paper, we analyze the tradeoff in the context of localization problems. In particular, the localization performance is characterized using the Cramer-Rao lower bound which reflects the performance limit in the family of unbiased estimators. For a variety of distance-sensitive sensing models, the bound comes in closed-form. In view of the tradeoff, we design a circular incremental inclusion scheme for efficient sensor tasking and propose to use the theoretical Cramer-Rao lower bound as a guiding heuristic. Our results reveal that there is a performance limit for collaborative sensing, and the expansion of sensing group size incurs fast diminishing returns on performance. The study results in new insights and provides general guidelines for sensor tasking in practical applications.