Intelligent traffic system, with which we have trained autonomous cars

Traffic AI™ is an external library which after integration with a simulator or video game, places cars on virtual roads that behave like in the real world. The system was used for teaching intelligent driving algorithms intended to control real, autonomous vehicles. We use Traffic AI™ in both professional simulators (training of ADAS driver assistance systems) and simulation games we create.

Independent library

Traffic AI™ is an advanced system which can be integrated with your project conveniently through API. It can run on a standalone machine for complex simulations or share hardware with your game or simulator.

Multiplaftorm integration

Traffic AI™ works very well with a wide range of modern game and simulation engines: Unity3D, Unigine, Unreal Engine 4, Cryengine and VBS. It is also ready to be integrated with your individual ideas and we are ready to fully help you in this process.

Intuitive configuration

Instead of dealing with the motion software of each car separately, you can set simulation parameters once and place intelligent vehicles on all roads simultaneously. Your system can take control over one of them at any time.

Intelligent traffic

Our car behavior models are based on intelligent algorithms – road traffic behaves like in the real world. Thanks to the possibility of additional configuration, your game or simulator will gain a completely new level of realism.

Traffic AI™ is an external library ready for integration with your project. Our experienced team can also do this integration for you. We communicate with your engine via API, process the received information about the road network, add the parameters and calculate the vehicle traffic simulation, to finally pass it back to your system.
The main functions of Traffic AI™ are:

  • Three types of vehicles: passenger cars, two-wheeled vehicles and emergency vehicles.
  • Dynamic behavior: lane changing, overtaking, cornering and collision avoidance.
  • A wide range of options: number of agents on the road, driving variables, behavior zones, etc.
  • Configuration of scenarios allowing you to plan specific training actions.
  • Creating a road grid based on the provided data and the use of multi-level roads.

Request a demo