Stationary environment models for Advanced Driver Assistance Systems

The vision of a fully self-driving autonomous car is year by year closer to be achieved. The crucial part of this vision is the perception which is carried by complex algorithms processing signals coming from a set of different sensors. A self-driving car has to know locations of stationary obstacles in its surrounding. This paper is an overview of existing models used to describe the stationary environment. Grid models like 1D, 2D or 3D discrete maps, primitive structures and free space boundary contours are described with some illustrative examples. Models are compared from the point of view of description completeness as well as applications. Estimates of memory consumption are also given.