We develop a method to detect atypical traffic jams in the City of Boston. Our motivation is to detect these traffic jams which are often caused by some event (e.g., accident, lane closure, etc.) and enable the City to intervene before congestion spreads and adjacent roads are negatively affected. Using a traffic jam dataset provided by the City of Boston, we present a novel detection system for anomalous jam identification. We demonstrate its effectiveness by using it to identify traffic jams that cannot be explained by typical traffic patterns.
Anomaly detection in transportation networks using machine learning techniques
2017-11-01
2087114 byte
Conference paper
Electronic Resource
English
Automatic Anomaly Detection with Machine Learning
NTRS | 2019
|Using Machine Learning for Advanced Anomaly Detection and Classification
British Library Conference Proceedings | 2016
|