Traffic estimation by fusing static and moving observations in highway networks
OPEN ACCESS
Date
2020-05
Publication Type
Conference Paper
ETH Bibliography
yes
Citations
Altmetric
OPEN ACCESS
Data
Rights / License
Abstract
Traffic monitoring and control constitute necessary steps in order to ensure the efficient function of transport networks. To that end, traffic state estimation and prediction are crucial tasks, normally relying on the (limited) deployment of sensor infrastructure. In recent years, apart from the traditional stationary sensors (i.e. loop detectors), new data collection alternatives have emerged at different levels of spatial andtemporal aggregation (e.g. Global Navigation Satellite System (GNSS) data, Bluetooth tracking, Car-floating data, etc.). The abundance of new data sources provides a unique opportunity to improve existing traffic monitoring strategies or develop new ones. A key role within this process is data fusion, the process of integrating multiple data sources in order to produce consistent and accurate state estimations. Furthermore, it is essential that the development of these techniques could be robust to data outages, among the various sources. In this regard, different data fusion techniques have been developed to allow an integration to take place. Transport networks are known for highly nonlinear behavior, posing a challenge and an opportunity with regard to the aspects above. This research deals with the issue of traffic state estimation for highway networks with limited, and potentially incomplete, measurement data from different sensor infrastructures. In the current framework, we suggest the use of Unscented Kalman Filter (UKF) which inherently incorporates the data fusion process in an algorithmic concept. In particular, we combine the second order traffic flow model METANET, with filtering methods (e.g. Unscented Kalman filter) properly modified to account for spatio-temporal correlations between the corresponding noise terms. Our results are compared against reference data to help us make decisive statements about the efficacy of the new methodology in tackling this problem.
Permanent link
Publication status
published
External links
Editor
Book title
Journal / series
Volume
Pages / Article No.
Publisher
STRC
Event
20th Swiss Transport Research Conference (STRC 2020) (virtual)
Edition / version
Methods
Software
Geographic location
Date collected
Date created
Subject
State estimation; Feedback control; Data-driven traffic assimilation; Highway road networks; Unscented Kalman filter; Data fusion; Cell transmission model (CTM); METANET model
Organisational unit
08686 - Gruppe Strassenverkehrstechnik
02655 - Netzwerk Stadt u. Landschaft ARCH u BAUG / Network City and Landscape ARCH and BAUG
Notes
Due to the Corona virus (COVID-19) the conference was conducted virtually.