Information Technology Reference
In-Depth Information
prognosis of problems and undesired situations, the determination of su-
cient dispatching actions and finally the propagation and realization of these
actions. At the end, the realization of actions is monitored by train move-
ment detection again. This process cycle is shown in Figure 1, which originally
was introduced in [1] to illustrate the bordering conditions within which an
asynchronous dispatching algorithm was evaluated.
Fig. 1. Integration and functionalities of computer based dispatching systems.
The different functionalities and components can be introduced shortly:
- Published timetable and basic data structures The timetable
is the central element for reliable, ecient and predictable train opera-
tion and the core data structure for a dispatching system: The system
starts with the published timetable
tt
orig
=
dp
0
. The forecasted train
runs are based on this timetable, which is modified successively due to
detected train movements, their reported positions and dispatching de-
cisions made. Each modification transforms
dp
i
into another timetable
dp
i
+1
, often called the
dispatching plan
. All operations are always ap-
plied to the last valid dispatching plan.
- Train position detection A correct and reliable detection of train
positions and the derivation of train movements is a requisite for dis-
patching systems. Different solutions are possible to enable discrete or
continuous detection. The detection based on discrete position messages
may be the most common approach, because systems with discrete ap-
proaches are widely available and standardized position telegrams are
specified by the UIC [2].
These telegrams encapsulate a time
t
,atrain
tr
and a position
p
. With
this information the route
r
orig
of
tr
is derived from the current dispatch-
ing plan
dp
i
.If
p
is found within
r
orig
the train trajectory is adjusted in
p
at
t
. Otherwise a new route
r
new
containing
p
must be determined
and a new trajectory is calculated. The new route
could
be the current
one, the corresponding trajectory again is adjusted in
p
at
t
. A successful
adjustment transforms
dp
i
into
dp
i
+1
containing the adjusted trajectory.
