Accuracy and Delay: An Inherent Trade-off in Cooperative UWB Navigation

Licentiate thesis, 2013

Location-aware applications and wireless sensor networks are becoming essential in our daily lives from a commercial, and public perspectives. The need of localization information to drive the applications is a key requirement. New technologies have emerged to tackle the problem of the limitations of the Global Positioning System (GPS) solutions. Ultra-wideband (UWB) is one of those emerging RF-technologies. The thrive in search for better accuracy involves improved ranging algorithms, higher transmission powers, and the use of cooperation among nodes. The goal of this thesis is to investigate the trade-off between medium access control (MAC) delay and accuracy for UWB systems based on hands-on experience and practical implementation with state-of-the-art equipment, based on two-way-ranging and a spatial time division multiple access scheme (STDMA). Paper A investigates the connection between accuracy and MAC delay for noncooperative scenarios. We quantify, by means of lower bounds how traditional methods to improve accuracy such as increased number of anchors, and increased communication range comes at a significant cost in terms of delay. Techniques such as selective ranging and eavesdropping help alleviate the trade-off and reduce the MAC delay in favor of mobile networks with tolerable accuracies. Paper B extends the work for cooperative scenarios, where nodes cooperate with each other by means of shared information. This sharing has an impact not only on the position accuracy but also on the MAC delay which we quantify by means of lower bounds, both for the accuracy and MAC delay. Once again, selective ranging is evaluated to reduce the MAC delay for finite cooperative networks. We show how indiscriminate cooperation leads to large MAC delays, which has a direct impact on the update rate for high mobility scenarios. Finally, Paper C unifies all findings by including derivations of the accuracy and MAC delay lower bounds for noncooperative and cooperative networks, evaluating selective ranging and eavesdropping to cope with the tradeoff in different conditions. Numerical evaluations are included for several distinct operations. Furthermore, we characterize the trade-off behavior for dense-location aware networks for both noncooperative and cooperative cases by means of scaling laws. We conclude by introducing a delay/accuracy parameter which can uniquely quantify the trade off between accuracy and MAC delay as a function of the agent and anchor density. Noncooperative eavesdropping shows to outperform cooperative networks in terms of accuracy with reasonable delays. Finally, in terms of scaling, we found that, under certain conditions, standard cooperative positioning exhibits the worst possible trade-off among the considered strategies.