AtEgoSim – Automated Generation of Highly Detailed Worlds, Situations, and AI Dynamics for Real-Time Research Driving Simulators

Guided by our experience with former versions of the HFC driving simulator, the AtEgoSim framework is a complete redesign of the simulator infrastructure with today’s technology. Its ambituous aim is to deliver a driving simulator tailored to the needs of psychological experts to generate a spectrum of well-controlled, repeatable traffic situations in a rich and varied virtual traffic environment in order to gain better understanding of the human behavior while performing a certain traffic-related task.

ategosimIn order to be amenable for the mainly non-technically trained domain specialists and leverage the complexity of the underlying simulator infrastructure, the AtEgoSim project’s key design goal is to ensure an ease of extensibility and adaptibility to the requirements of the next driving experiments at hand. The past version of the HFC driving facility suffered from requiring a core developer to add new behaviours and scenarios, which was both time- and cost-sensitive.

To this end, the AtEgoSim project devises a GUI-frontend that makes it easy for the user to manage and extend scenarios through custom behaviours and orchestrated screen-plays by facilitating a visual scripting interface. More advanced users can take advantage of a full-blown scripting language.

Another important feature of an immersive real time driving simulator is how credibly machine-controlled traffic participants behave: participants driving in a synthetically clean manner or, even worse, cars locomoting in a physically disputable way disrupt the realism of the whole experience. To this end, the AtEgoSim project strictly separates between human-like controllers and complex, physically based vehicle models, whose dynamics can be optionally calculated by dedicated external software. For the sake of abstraction in scripted screen-plays, the project strives to strike the delicate balance between fixed-function low-level behaviour of machine-controlled traffic participants and higher-level behaviours the user is really only interested in adapting.

The AtEgoSim project is a collaboration, funded by the AiF, between the two industrial partners TrianGraphics GmbH and HFC Human-Factors-Consult GmbH and the University of Bonn, Germany, Department of Computer Graphics.

Competencies

Services

  • Requirements Definition
  • User Interfaces
  • GUI-Design