CANOE: Opportunistic calibration assisted micro navigation in dense open environments

Abstract

Crowded musical concerts (and open densely crowded environments in general) pose significant hurdles to people trying to navigate to their family and friends. Calls get frequently dropped due to crowd spikes and loud noise makes cellular voice communication cumbersome. Low visibility in such dense environments render traditional navigation solutions ineffective. Global Positioning System (GPS) and Pedestrian Dead Reckoning (PDR) based systems are known to be error prone and inefficient for such micro-navigation scenarios. Incorrect and delayed position fixes result in high convergence time and frequent oscillating about the route to the destination, leading to a frustrating user experience, specially in dense crowd. To address the dual issue of convergence and user experience we propose CANOE - a novel tunnel based navigation methodology which allows users the flexibility of using their own sense to wade through crowd, while concurrently using best effort opportunistic position fixes to constrain drifts. It also reduces the dependency of constantly looking at the phone for position updates and route information. Extensive simulation results show that our algorithm can achieve 2-4x improvement in convergence time and reduced oscillations under different crowd scenarios as compared to state-of-the-art approaches. We also conduct live experiments with a crowd of 22 people and 15 smartphones and find that CANOE can contain location drifts within 1-2m in signal degraded environments where PDR location drifts range around 6-7m for short walks of 25m.