Collaborative Effects of Mobile Sensor Network Localization Through Distributed Multimodal Navigation Sensor Fusion

Abstract

This paper addresses the problem of localizing both absolute and relative 3-D positions of mobile sensor network nodes by fusing inertial measurements and radio frequency (RF) ranging measurements. Four online estimation algorithms are described: centralized optimization, distributed optimization, centralized Extended Kalman Filtering (EKF), and distributed EKF. We derive an error scaling law, termed the “Teamwork Effect, epsilon (n,s) oc epsilon(1,s)/ (division symbol) n, where epsilon(1,s) is the error of each node’s individual location estimate as a function of s (e.g. time or distance traveled) and epsilon(n,s) is the improved location error a node can expect when collaborating as a member of a size n network, under idealized assumptions. Analysis and simulations in a variety of scenarios show that this law holds even when the idealized assumptions are violated and that all four algorithmic approaches can localize both the absolute and relative positions of a mobile network with high accuracy, even with greatly varying INU and RF sensor noise models. By comparing the trade off between performance and computational complexities, we conclude that the distributed EKF algorithm is a good candidate for realtime system implementation.”