Inquiry on Recent Progress

[Xiao Li]

Hi All

I'm just wondering how this project is progressing since the last post is almost a year ago. I'm trying to implement a cooperative mapping project using multiple robots and would like to be able to deploy distributed controllers on each robot. Therefore it would be nice if I can have ROS running independently on each robot. The multimaster wiki only goes up to groovy and there's little documentation on actual implementation. If some up updates can be provided that would be great. Thank you!

Xiao

[Alexander Tiderko]

Hi Xiao,

there are currently two working and maintained multimaster implementations:

rocon: http://wiki.ros.org/rocon

multimater_fkie: http://wiki.ros.org/multimaster_fkie

try it out!

Regards,

Alex

[Daniel Stonier]

We've been quiet for a little while, mostly because we're working on hard on the framework to get it to the point where it can be really useful and stable for a while (i.e. like a navigation stack) for people to start doing their own thing on top without it changing too frequently . You can get an idea of where we are going with it on our development wiki:

• http://redmine.robotconcert.org/projects/opp/wiki
• http://redmine.robotconcert.org/projects/opp/wiki/Milestones
  

The folk at the University of Texas at Austin have been great in giving us an extra use case and providing us with some valuable input on getting the framework designed in such a way that we supply most of the peripheral requirements, but leave them free to build their higher level software in whatever way they wish (they're using orc, we've another korean group trying bpel). They're also going to implement fully a prototyped Gazebo service into the multimaster framework so that you can easily test/develop without the robots. OSRF also has been helping build a decent layer of abstraction inside the robots (capability) allowing us to write robot apps more portably.

I'm planning to do a release of some of the tools we have built that are not multimaster-centric in the next fortnight or so, but for a full release with documentation we're probably looking at june-indigo.

The big jobs over this time period include gazebo, software farm, a map/world annotation module and investigation into the tf problem. 

If you're interested in what's going on/or would like to provide some input, I'd be happy to have a hangout or further discussion here.

Cheers,

Daniel.

  

[Xiao Li]

Alex

Thanks for the links!

[Xiao Li]

Daniel

Also thank you for this detailed update. I'd definitely like to get more involved and help in any way I can. It looks like you are currently developing the features of ROCON in various repositories and I can't seem to find a systematic description of how to use the packages (or maybe I'm not looking at the right places). Does this mean I'd have wait for your more official release in Indigo?

Xiao 

[Daniel Stonier]

On 1 March 2014 04:15, Xiao Li <xli...@gmail.com> wrote:

Daniel

Also thank you for this detailed update. I'd definitely like to get more involved and help in any way I can. It looks like you are currently developing the features of ROCON in various repositories and I can't seem to find a systematic description of how to use the packages (or maybe I'm not looking at the right places). Does this mean I'd have wait for your more official release in Indigo?

Xiao,

Sorry for the delay, last week got rampant on me. We're actively developing on the hydro-devel branches right now but won't release or update the documentation for hydro since that would affect the current turtlebot release which uses some of it.

Unfortunately that means I don't have too much tutorial style documentation lying around. However you can start with a rosinstall - you'll always find the required rosinstalls in the relevant branch of the https://github.com/robotics-in-concert/rocon repo. Right now I'm developing with

• hydro-devel/rocon.rosinstall on ros hydro/precise and
  
• indigo/rocon.rosinstall on ros indigo/saucy
  

The sources for both are still compatible though - I merge anything from indigo back into hydro-devel. That should give you a working source workspace. 

To play around - you might like to go through things in this order:

• rocon_interactions : it can be used outside of multimaster, but it's also very important for us to connect users to a running concert system. The indigo tutorials also work for the hydro-devel branch.
  
• The key enabler for multimaster is the rocon gateway/hub system. The rocon gateway tutorials are unchanged, so you can follow them. The only changes we've made to this system is to bugfix and have done a partial implementation on wireless handling.
  
• The rocon tutorials chatter_concert and turtle_concert are still there but utilise a very different higher level framework for services and scheduling. You should be able to reverse engineer what is going on with these.

The documents scribbled on our developer's redmine can give you a better feeling for what is going on right now. 

If you find bugs, hit us on github or if you want to ask questions, either here or to me directly is fine (I'll be alot more responsive ;)). If you'd like to have a hangout this week or next sometime that'd be great - I'd like to hear what your use case is about.

Cheers,

Daniel.

 

Xiao 

On Thursday, February 27, 2014 2:01:42 AM UTC-5, Daniel Stonier wrote:

On 27 February 2014 11:07, Xiao Li <xli...@gmail.com> wrote:

Hi All

I'm just wondering how this project is progressing since the last post is almost a year ago. I'm trying to implement a cooperative mapping project using multiple robots and would like to be able to deploy distributed controllers on each robot. Therefore it would be nice if I can have ROS running independently on each robot. The multimaster wiki only goes up to groovy and there's little documentation on actual implementation. If some up updates can be provided that would be great. Thank you!

Xiao

We've been quiet for a little while, mostly because we're working on hard on the framework to get it to the point where it can be really useful and stable for a while (i.e. like a navigation stack) for people to start doing their own thing on top without it changing too frequently . You can get an idea of where we are going with it on our development wiki:

• http://redmine.robotconcert.org/projects/opp/wiki
• http://redmine.robotconcert.org/projects/opp/wiki/Milestones
  

The folk at the University of Texas at Austin have been great in giving us an extra use case and providing us with some valuable input on getting the framework designed in such a way that we supply most of the peripheral requirements, but leave them free to build their higher level software in whatever way they wish (they're using orc, we've another korean group trying bpel). They're also going to implement fully a prototyped Gazebo service into the multimaster framework so that you can easily test/develop without the robots. OSRF also has been helping build a decent layer of abstraction inside the robots (capability) allowing us to write robot apps more portably.

I'm planning to do a release of some of the tools we have built that are not multimaster-centric in the next fortnight or so, but for a full release with documentation we're probably looking at june-indigo.

The big jobs over this time period include gazebo, software farm, a map/world annotation module and investigation into the tf problem. 

If you're interested in what's going on/or would like to provide some input, I'd be happy to have a hangout or further discussion here.

Cheers,

Daniel.

  

--
You received this message because you are subscribed to the Google Groups "ROS Multi-Master Special Interest Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ros-sig-mm+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
Phone : +82-10-5400-3296 (010-5400-3296)
Home: http://snorriheim.dnsdojo.com/

Yujin R&D: http://rnd.yujinrobot.com/

[Vinícius Alves Hax]

Hello guys, I am a master student from Brasil and I am working on a
distributed naming resolution baseed on DHT for ROS compatible with
ROS Hydro.

There is some code here for the project
https://github.com/viniciusah/ros_mmaster but I made some changes on
roscpp too that are not on repository.

My results are not good yet (time for name resolution is worse that on
ROS Hydro for now), but when I achieve better results I share them
with you if someone wish.

Thanks and best wishes!

> For more options, visit https://groups.google.com/d/optout.

[Daniel Stonier]

On 15 March 2014 00:05, Vinícius Alves Hax <vinic...@gmail.com> wrote:

Hello guys, I am a master student from Brasil and I am working on a
distributed naming resolution baseed on DHT for ROS compatible with
ROS Hydro.

There is some code here for the project
https://github.com/viniciusah/ros_mmaster but I made some changes on
roscpp too that are not on repository.

My results are not good yet (time for name resolution is worse that on
ROS Hydro for now), but when I achieve better results I share them
with you if someone wish.

Thanks and best wishes!

I missed this earlier. DHT is an interesting idea - how did you fare with it?

 

Daniel.

[Vinícius Alves Hax]

Hello Daniel and everyone,

I make some tests on last week and the naming resolution is up to 3
times faster in my implementation (tested on a LAN network).

I am almost finishing my dissertation text. In the next week I will
made a blog post with the summary of my research and I send the link
for the group list.

[]s