[Back]


Talks and Poster Presentations (with Proceedings-Entry):

T. Blazek, M. Gasser, C. Mecklenbräuker:
"Modeling Vehicle Distributions at Urban Intersections using the Information Bottleneck";
Talk: Dependable Wireless Communications and Localization for the IoT, Graz; 09-12-2017; in: "Dependable Wireless Communications and Localization for the IoT", (2017), 3 pages.



English abstract:
Accurate simulations of Vehicle-to-Any communi- cation networks at urban intersections need to be based on precise models. We approach this by analyzing the distribution of neighboring vehicles a communication node sees as a function of the distance to the closest intersection. Our analysis is based on real world map data from the city of Linz in Austria, and used both static random placement of vehicles, as well as the output of a vehicular traffic simulator (SUMO). Using the Information Bottleneck method, we discretize the road into a set amount of intervals, and provide cumulative distribution functions for these intervals. Through evaluating the resulting data for different vehicle densities, we show that a low-complexity discretization for the model of neighboring vehicles performs well. As a second result, we demonstrate strongly diverging behavior between static placement and SUMO simulations, proving the necessity for accurate model assumptions. In this work, we provide the optimal values of the interval boundaries, as well as the cumulative distribution functions of the neighbors for all intervals.

German abstract:
Accurate simulations of Vehicle-to-Any communi- cation networks at urban intersections need to be based on precise models. We approach this by analyzing the distribution of neighboring vehicles a communication node sees as a function of the distance to the closest intersection. Our analysis is based on real world map data from the city of Linz in Austria, and used both static random placement of vehicles, as well as the output of a vehicular traffic simulator (SUMO). Using the Information Bottleneck method, we discretize the road into a set amount of intervals, and provide cumulative distribution functions for these intervals. Through evaluating the resulting data for different vehicle densities, we show that a low-complexity discretization for the model of neighboring vehicles performs well. As a second result, we demonstrate strongly diverging behavior between static placement and SUMO simulations, proving the necessity for accurate model assumptions. In this work, we provide the optimal values of the interval boundaries, as well as the cumulative distribution functions of the neighbors for all intervals.

Keywords:
Information Bottleneck; Vehicle Distributions


Electronic version of the publication:
http://publik.tuwien.ac.at/files/publik_261347.pdf


Created from the Publication Database of the Vienna University of Technology.