__TORCS.txt

At the car level, new simulation modules can be developed,
which include intelligent control systems for various car components. At the
driver level, a low-level API gives detailed (but only partial) access to the simulation
state. This could be used to develop anything from mid-level control
systems to complex driving agents that find optimal racing lines, react successfully
in unexpected situations and make good tactical race decisions. Finally,
for researchers that like a challenge and are also interested in visual processing,
a 3d projection interface is available

In TORCS, the participating players are referred to as “robots”. They are
loaded as external modules in TORCS. This means that new artificially intelligent
agents can be developed independently and they only have to satisfy the
basic API requirements for robot code

Before every race, each robot can gather and process information about the
track’s geometry and surfaces. It is up to the user to decide how much of this
information use

robots have 50FPS

The default interface is through a low-level API2 which can provide detailed
information about the race status to the robot, exact position, distance from the
edge of the track, the position of other cars, etc. However, there are many parts
of the simulation state to which the robots have no direct access. Consequently,
even the basic driving problem is partially observable

Robots may also use a calculated three-dimensional projection, instead of
the low-level API. This is intended to be used for researchers that have an
interest in visual processing. However, the overall problem then becomes much
harder as there is significantly less information directly available

The racing problem could be split into a number of different components, including
robust control of the vehicle, dynamic and static trajectory planning,
car setup, inference and vision, tactical decisions (such as overtaking) and fi-
nally overall racing strategy. With only a single car on the track, the overall
problem can be formalised as a partially observable Markov decision processes.

!!! mit nur einem Auto ist es nen POMDP!!!