Trafﬁc Congestion Estimation Using Floating Car Data (External thesis / BMW) (Diplomathesis)
21.03.2014, 14:00, room 3945
This diploma thesis investigates the estimation and prediction of trafﬁc congestion using ﬂoating car data. Unlike previous works dealing with infrastructure-based trafﬁc density estimation, this thesis focuses on estimating the density based only on speed and distance to the preceding vehicle. Input data was collected within the simTD project.
A trafﬁc simulation tool SUMO has been used for testing the methods before deploying them on real data. For this purpose, various approaches and methods namely fuzzy system and particle ﬁlter were tested. The number of probe vehicles within the trafﬁc inﬂuences the accuracy of results.
The higher the number of probe cars, the more robust the estimation results. This work examines the sufﬁcient percentage of probe cars within the total trafﬁc in order to get accurate trafﬁc state evaluation.