Fine-Grained Stop-Move Detection with UWB: Quality Metrics and Real-World Evaluation

The synergy between the accurate trajectories offered by ultra-wideband (UWB) systems and techniques to extract higher-level mobility patterns is largely unexplored. We study whether staple techniques designed for systems with coarser resolution apply to UWB, investigating quantitatively the quality of the fine-grained analyses enabled by the latter. To this end, we contribute a novel family of metrics suited to the high UWB spatio-temporal resolution and use them to configure and ascertain the quality of representative techniques along several dimensions. We focus on the well-known stop-move pattern and derive our findings from a real museum setting with the use case of capturing visits to exhibits. We acquire UWB trajectories in both controlled ( in vitro ) and uncontrolled ( in vivo ) conditions, along with ground truth. Despite exhibits being very close to each other, our results show that stops near them can be correctly identified and associated in the vast majority of cases and with very small spatio-temporal error. These positive results from real-world experiments, along with our technical contributions, open new opportunities in exploiting UWB for mobility analyses.