VehiLux model
The VehiLux model generates vehicular traces using freely accessible data sources, like open-source maps and real traffic counting data [1]. Parameters of the model were tuned by genetic algorithms (GAs) [2]. The model can be applied easily to any area for which map topology and basic traffic volume counting data exist. The overall architecture is presented in Fig. 1.
Fig 1. VehiLux model.
Implementation details are availible on VehiLux project page
Generated traces consider:
- spatial,
- temporal,
- and behavioural
aspects of vehicular traffic. Uneven distribution over a space is obtained by an original traffic demand model. It specifies individual origin-destination pairs (OD) in a process presented in Fig. 2.
Fig 2. Probability model.
A trace is additionally determined with a departure time that enables to reproduce the rush-hour peak in traffic demand. Traces are further assigned by a vehicle type that brings heterogeneity into microsimulation (e.g. acceleration and maximum speed capacities).
- References:
- Y. Pigne, G. Danoy, and P. Bouvry., "A vehicular mobility model based on real traffic counting data.", In Thomas Strang et al., editors, Nets4Cars/Nets4Trains 2011, Volume 6596, Series Lecture Notes in Computer Science, Pages 131-142, Springer, 2011
- M. Seredynski, G. Danoy, M. Tabatabaei, P. Bouvry, and Y. Pign ́e., "Generation of Realistic Mobility for VANETs Using Genetic Algorithms.", In Proc. of the IEEE Congress on Evolutionary Computation, pages 2A1–1–2A1–2, 2012 [doi]
