Question about project development

[Peque]

Hi,

I am trying to run the simulator from the particle_filter branch. I managed to compile everything but I do not seem to be able to load a maze into the Gazebo world, and it seems that information is missing in the README file. Is there another documentation source?

Also, I have seen you are currently working in other branches (i.e.: master, new_sim). Which are the differences with respect to particle_filter? Should I be using one of those instead or are they still under heavy development?

Thanks for your help.

[PeterMitrano]

the particle filter branch never went anywhere. try new_sim if you're interested in the simulator. But just to warn you everything is under heavy development and will often be broken

[Peque]

@PeterMitrano Thanks. I managed to compile the new_sim branch as well. Had to change 3 lines to make it work under Fedora 26:

diff --git a/sim/CMakeLists.txt b/sim/CMakeLists.txt
@@ -1,11 +1,11 @@
 #cmake_policy(SET CMP0071 OLD)

-find_package(Qt5 5.9.1 REQUIRED COMPONENTS Core Gui Charts)
+find_package(Qt5 5 REQUIRED COMPONENTS Core Gui Charts)
 find_package(Protobuf REQUIRED)
 find_package(ignition-msgs0 QUIET REQUIRED)
 find_package(ignition-transport3 QUIET REQUIRED)

-set(PROTOBUF_IMPORT_DIRS ${IGNITION-MSGS_PROTO_PATH} "/usr/include/")
+set(PROTOBUF_IMPORT_DIRS ${IGNITION-MSGS_PROTO_PATH} "/usr/include/" "/usr/include/ignition/msgs0/")

 add_definitions(${QT_DEFINITIONS})
 include_directories(${PROTOBUF_INCLUDE_DIRS} ${IGNITION-TRANSPORT_INCLUDE_DIRS})
diff --git a/sim/setup.sh b/sim/setup.sh
@@ -1,6 +1,6 @@
 # MUST BE RUN FROM root of the project (Smartmouse_2017)
-#source /usr/share/gazebo/setup.sh
-source /usr/local/share/gazebo/setup.sh
+source /usr/share/gazebo/setup.sh
+#source /usr/local/share/gazebo/setup.sh
 export GAZEBO_PLUGIN_PATH=${PWD}/.sim_build/sim:$GAZEBO_PLUGIN_PATH
 export LD_LIBRARY_PATH=${PWD}/.sim_build/sim:/usr/local/lib/$LD_LIBRARY_PATH
 export GAZEBO_MODEL_PATH=${PWD}/sim/models

I have a couple of questions/comments:

• Is it really necessary to make Qt 5.9 required? Is not 5.7 enough? (5.7 is the one available in the Fedora 26 official repositories)
• I am not sure why I had to add the "/usr/include/ignition/msgs0/" there. Do you have any idea?
• Would you mind adding instructions on how to launch the simulation with a sample maze in the README file?

[PeterMitrano]

The simulation isn't finished so I can't give any instructions. Maybe in a few months but not any time soon. What are you trying to use this for? There are probably other tools you should be using (gazebo).

[Peque]

@PeterMitrano I would like to have a micromouse simulation environment (see this issue in my open micromouse project). Even if the models are very rudimentary, just wanted to be able to try my ideas/code/algorithms before actually loading them in the micromouse. Also, I thought this could also help with testing, to make sure changes in the code will never make the micromouse fail while searching or solving for the shortest path (i.e.: with the real micromouse, at home, you probably do not have a full-size maze, but you could test big/complex mazes with the simulator).

I thought maybe this project would help as a good starting point and save me some time. Thought you were doing exactly that: a micromouse solving simulator with Gazebo. Is not that your idea?

[PeterMitrano]

That was the old system, and can be found in last year's repo + the gzmaze repo. This is me writing my own simulator exclusively for micromouse from scratch.

On Mon, Oct 2, 2017, 03:52 Miguel Sánchez de León Peque < notifications@github.com> wrote:

@PeterMitrano https://github.com/petermitrano I would like to have a micromouse simulation environment (see this issue in my micromouse project https://github.com/Theseus/theseus/issues/74). Even if the models are very rudimentary, just wanted to be able to try my ideas/code/algorithms before actually loading them in the micromouse. Also, I thought this could also help with testing, to make sure changes in the code will never make the micromouse fail while searching or solving for the shortest path.

I thought maybe this project would help as a good starting point and save me some time. Thought you were doing exactly that: a micromouse solving simulator with Gazebo. Is not that your idea?

— You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub https://github.com/WPISmartmouse/SmartMouse_2018/issues/1#issuecomment-333464010, or mute the thread https://github.com/notifications/unsubscribe-auth/AD0zEkLOe40V4megQ2OfqXxLfjCa0Tg0ks5soJZVgaJpZM4PpvsA .

[Peque]

@PeterMitrano Out of curiosity, are you making this for fun/learn or did you actually find some limitations with the Gazebo simulator? I am asking because I was thinking about creating a Gazebo-based micromouse simulator just like you, and want to know if I will end up finding the same limitations (if they exist). Thanks for your attention.

[PeterMitrano]

In making it because gazebo is too complex and too slow. It works and if you can learn how it works it can do amazing things, but the physics are too complex for my needs. Also, I never got more then 4x real time simulation but my new simulator can do >10x

On Tue, Oct 3, 2017, 12:24 Miguel Sánchez de León Peque < notifications@github.com> wrote:

@PeterMitrano https://github.com/petermitrano Out of curiosity, are you making this for fun/learn or did you actually find some limitations with the Gazebo simulator? I am asking because I was thinking about creating a Gazebo-based micromouse simulator just like you, and want to know if I will end up finding the same limitations (if they exist). Thanks for your attention.

— You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub https://github.com/WPISmartmouse/SmartMouse_2018/issues/1#issuecomment-333896067, or mute the thread https://github.com/notifications/unsubscribe-auth/AD0zEqqTiu8vLBsWqirFETvB1p1B8_c5ks5sol_CgaJpZM4PpvsA .