Driving simulation is widely used by automotive companies as well as road traffic, training and safety research laboratories for efficient vehicle engineering design, driver behavior studies, better road safety and driver training. Though a number of industrial software solutions are available on the market such as SimVehicleLT™, SCANeR™, STISIM Drive® or Virtual Test Drive, their deployment may be heavy, often requires specific training and adaptation for specific need can be lengthy and costly.
As an alternative, open source codes are free of charge and typically easy to adapt and/or modify, among them OpenSD2S. This recently introduced Open Scalable Driving Simulation Software has a modular and scalable architecture based on MPI (Message Passing Interface) which allows for distributing the computational load over a computer cluster. The simulator is split into as many modules as features: a car dynamics module based on ODE (Open Dynamics Engine), a visual module based on OGRE (Open source 3D GRaphics Engine) or OSG (Open Scene Graph) which allows for multi-screen support, a sound module based on OpenAL (Open Audio Library), among others.
To meet the requirements for driver behavior studies, a traffic module, based on the OpenDRIVE road format and a simulation data recorder have been implemented. Thanks to its flexibility and these new features, OpenSD2S is now used for road driver and traffic research experimentation, such as at University of Lyon, where driver cognitive load study is carried out during road exit maneuvers. The used hardware configuration is an affordable solution for such an experiment, using a mainstream gaming steering wheel, pedals and a standard driver seat.
Future work will be carried out to improve the vehicle dynamics realism (collisions and suspension support), tools for visual and road database generation as well as advanced traffic model. The currently existing embedded version of the vehicle dynamics will also be extended to an embedded OpenSD2S version.