Raw GIS to 3D road modeling for real-time traffic simulation
- PDF / 4,089,442 Bytes
- 18 Pages / 595.276 x 790.866 pts Page_size
- 101 Downloads / 185 Views
ORIGINAL ARTICLE
Raw GIS to 3D road modeling for real-time traffic simulation Yacine Amara1 · Abdenour Amamra1 · Salim Khemis1 Accepted: 30 October 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract In this work, we propose a new approach to road modeling and 3D traffic simulation. Based on the raw geographic information system (GIS) data laid out as sparse polylines with attributes, we compute a more adequate functional description for realtime simulation of on-road vehicle animation. The proposed approach begins with a filtering/subdivision module where the raw polylines are transformed into a graph of functional road segments as arcs and the nodes as intersections. Then, the vehicle speed profile is computed based on its dynamics, its neighborhood and the curvature profile of the road. Afterward, a multi-agent system is proposed in order to handle a large number of simulated vehicle/driver couples. Finally, we deploy a 3D rendering engine to display the computed 3D simulation on screen. The resulting model satisfies most of the real road features for traffic simulation including road interchanges, roundabouts, intersections, lanes, etc. More importantly, the simulated driving qualitatively mimics the real behavior of the drivers/vehicles on the road as can be seen in the accompanying video (RTSP video). We also validate our findings with a technical assessment based on macroscopic and microscopic traffic simulation metrics in several road traffic scenarios. Keywords 3D road traffic simulation · Road modeling · 3D traffic rendering · Vehicle virtual navigation
1 Introduction The availability of large amounts of geographic information data has motivated many researchers to model and simulate the phenomenon of road traffic for both research and commercial purposes. Nevertheless, good quality data are mostly owned by large corporations such as Google and Microsoft and kept locked away to maintain a commercial advantage. On the other hand, the freely available crowd-sourced data such as that of OpenStreetMap.1 are released in the rawest form that requires a huge amount of preprocessing ahead of utilization [1]. Road traffic simulation has become a necessary analysis tool for the study of phenomena related to road infrastructure congestion. This technology enables quantifying the impact of an action or an incident on road traffic fluidity without the actual deployment of the physical infrastructure. Such an operation requires the cooperation of all the intervening 1
www.openstreetmap.org
B 1
Abdenour Amamra [email protected]
agents, hence inducing significant costs and planning time. Besides, the use of 3D simulation techniques allows for a fast and flexible realistic reproduction of the same scenario at a minor expense. In this context, we focused on designing and implementing several algorithms for 3D simulation of road traffic all gathered in what we dubbed Road Traffic Simulation Platform (RTSP). This tool should fulfill some requirements such as the analysis of the impact of
Data Loading...
