Gábor Balázs
2006-09-11 08:18:12 UTC
Hello Oliver,
We collected a couple of questions about these webots versions.
We used 5.1.6 under windows xp (because 5.1.7 became unstable there
after the controller modification), and version 5.1.7 under gentoo linux.
1) In both versions the simulations are finished by the environment after
100 seconds (independently from battery state).
Why does it happen? Can you disable it?
2) In both versions if you press the Alt or the Ctrl button, the environment
thinks you never release them,
so you have to revert the simulation to get the control back. It is very
disturbing after an Alt+Tab pressing.
Anyway the robot_keyboard_get_key function hardly notices the keyboard
events if we use a few (more than 3)
robot_console_printf function call in every simulation step (we only
tested it with the 5.1.6 version).
3) In 5.1.7 version after the simulation finishes I think the robot_die
function is not called because the processes are killed
and the appropriate string is not printed on the console. It works fine
in the 5.1.6 version.
4) The webots internal text editor is always opened after a simulation
revert.
Can I disable it somehow? :)
5) In the 5.1.7 version under linux, it is printed sometimes:
transform field # = 0x3
** ERROR **: file api/scene_manager.c: line 140
(scene_manager_read_transform_update): should not be reached
aborting...
Anyway I don't know when exactly, because I didn't notice any problem in
the environment because of it.
6) I noticed that Aibos' don't have GPS. This info would be good for us
later in fitness functions, so I tried to
"simulate" GPS by the supervisor function, but I have two questions
about the emitter - reciever concept:
- How you can set the channel of a receiver like in the case of
emitter with emitter_set channel?
- What are the receiver_get_x and receiver_get_z functions?
(these seem to be what is needed for GPS, but I don't know how
to use them
and they are not in the API documentation)
7) Is that possible that the supervisor process pause the simulation?
For example to run a new thread from a robot controller which makes some
computation until the simulation is paused.
Sorry for the lot of questions. :)
Br,
`bg` & cy
We collected a couple of questions about these webots versions.
We used 5.1.6 under windows xp (because 5.1.7 became unstable there
after the controller modification), and version 5.1.7 under gentoo linux.
1) In both versions the simulations are finished by the environment after
100 seconds (independently from battery state).
Why does it happen? Can you disable it?
2) In both versions if you press the Alt or the Ctrl button, the environment
thinks you never release them,
so you have to revert the simulation to get the control back. It is very
disturbing after an Alt+Tab pressing.
Anyway the robot_keyboard_get_key function hardly notices the keyboard
events if we use a few (more than 3)
robot_console_printf function call in every simulation step (we only
tested it with the 5.1.6 version).
3) In 5.1.7 version after the simulation finishes I think the robot_die
function is not called because the processes are killed
and the appropriate string is not printed on the console. It works fine
in the 5.1.6 version.
4) The webots internal text editor is always opened after a simulation
revert.
Can I disable it somehow? :)
5) In the 5.1.7 version under linux, it is printed sometimes:
transform field # = 0x3
** ERROR **: file api/scene_manager.c: line 140
(scene_manager_read_transform_update): should not be reached
aborting...
Anyway I don't know when exactly, because I didn't notice any problem in
the environment because of it.
6) I noticed that Aibos' don't have GPS. This info would be good for us
later in fitness functions, so I tried to
"simulate" GPS by the supervisor function, but I have two questions
about the emitter - reciever concept:
- How you can set the channel of a receiver like in the case of
emitter with emitter_set channel?
- What are the receiver_get_x and receiver_get_z functions?
(these seem to be what is needed for GPS, but I don't know how
to use them
and they are not in the API documentation)
7) Is that possible that the supervisor process pause the simulation?
For example to run a new thread from a robot controller which makes some
computation until the simulation is paused.
Sorry for the lot of questions. :)
Br,
`bg` & cy