In model-based vehicle tracking a previously learned 3D model of a vehicle is used to detect the desired vehicle in a camera image. 
The overall task is to estimate the position and the orientation of the leading vehicle.
Hypothesis for possible positions are evaluated based on detected features.

These features are:

  • edge features (e.g. car windows)
  • area features (e.g. red areas)
  • point features (e.g. surf features)
  • heat features (e.g. heat of the exhaust)

This system is used for autonomous convoy driving, where the autonomous vehicle follows the leading vehicle.