Information Technology Reference
In-Depth Information
Fig. 6.
State variables and resources describing the robotic platform and the communication chan-
nel availability
Fig. 7.
Execution time (in milliseconds) for the GOAC-
plus
domain (for increasing number of
TakingPicture
goals)
Given an initial situation in which the robot is at location
0
,
0
and the pan-tilt is ori-
ented to
, we scale the problem by adding different
TakingPicture
goals.
Figure 7 shows execution time (in milliseconds) of our benchmark problem (the number
of
TakingPicture
goals on the abscissas) for all the presented planners. We can observe
how the J-
TRE
(stp)
planner scales quite well with respect to all the problem instances.
On the contrary, the intrinsic complexity of the problem overwhelms both the J-
TRE
(ac)
and the J-
TRE
(smt)
planners that only solve the single smaller instance of the problem.
0
,
0
6
Conclusions
Temporal flexibility required in controlling mechanisms in real-time (i.e., robotics),
interacting with agents requirements as well as uncertainty of real world domains, are