A Bayesian approach to detect pedestrian destination-sequences from WiFi signatures
Abstract
WiFi traces are used to detect pedestrian activity-episode sequence. They are merged with other data: attractivity, map and time constraints. Ambiguity of the data is explicitly modeled. Novel method to generate candidate activity-episode sequences and their likelihood. Detailed validation and sensitivity analysis is performed.
- Publication:
-
Transportation Research Part C: Emerging Technologies
- Pub Date:
- July 2014
- DOI:
- 10.1016/j.trc.2014.03.015
- Bibcode:
- 2014TRPC...44..146D
- Keywords:
-
- Network traces;
- Activity choice modeling;
- Pedestrians;
- Semantically-enriched routing graph (SERG);
- Potential attractivity measure;
- Activity-episode sequence