The research was based on datasets collected at MUSE (the science museum of Trento, Italy) and concerned three different aspects.

1. Quantitative evaluation of state-of-the-art stop detection methods: we defined a family of metrics to compare the performance with respect to ground truth collected at MUSE, used in different experimental contexts to evaluate the effectiveness of the methods, their robustness and their optimal configuration.

2. Realization of the pipeline to transform the UWB trajectories into semantic trajectories, defined at a higher level of abstraction, in order to simplify the data analysis.

3. Analysis and interpretation of the data, for the use of the museum curators, to evaluate the effectiveness of the paths and the exhibition, e.g., to determine which of the exhibited objects are of greater interest to the visitors. Together, these define an innovative methodology for the analysis of stops at high spatial-temporal resolution. The activity has generated numerous datasets; the largest one contains 1500 trajectories of real visitors, who participated with informed consent and in total anonymity.