Which version of ros_control to use for new designs

[Mart Moerdijk]

Hi all,

I'm currently looking into implementing a hardware interface for ros_control. On the github[1] page of ros-controls there are two implementations, ros_control and ros_control2. Could someone explain to me what the difference is between them? And witch version I should use if I want to implement a new hardware interface today? 

Thanks,

Mart

[1] https://github.com/ros-controls

[G.A. vd. Hoorn - 3ME]

On 27-10-2017 11:12, Mart Moerdijk wrote:
> Hi all,
>
> I'm currently looking into implementing a hardware interface for
> ros_control. On the github[1] page of ros-controls there are two
> implementations, ros_control and ros_control2. Could someone explain to me
> what the difference is between them? And witch version I should use if I
> want to implement a new hardware interface today?

ros_control2 was presented at ROSCon16 in Vancouver, and is a new
implementation exploiting the features that ROS 2 brings.

If you're targetting ROS 1, you'd use 'ros_control'.

For ROS 2, 'ros_control2' might make sense.


Gijs


> [1] https://github.com/ros-controls

[Bence Magyar]

Hiya, 

I would also add that ros_control2 has only been developed as an early proof of concept and is - as most parts of ROS2 -  far from final. We don't have a proper hardware implementation yet. 

Fix: Roscon 2017 ;) 

Cheers,

Bence

[1] https://github.com/ros-controls

--
You received this message because you are subscribed to the Google Groups "ROS/Orocos Robot Control Special Interest Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ros-sig-robot-control+unsubscri...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[G.A. vd. Hoorn - 3ME]

On 27-10-2017 15:38, Bence Magyar wrote:
> Hiya,
>
> I would also add that ros_control2 has only been developed as an early
> proof of concept and is - as most parts of ROS2 - far from final. We don't
> have a proper hardware implementation yet.
>
> Fix: Roscon 2017 ;)

hm, yes :)


Gijs

[Mart Moerdijk]

Thanks all for you replies.  I was there at roscon 2017 and actually heard the talk, but wenn I saw the ros_control2 on github I did not directly link it to ros2. I just had a flashback to me implementing a IRCv1 although would have been better to implement the IRCv2 for new designs, but that was not documented :-) 

--
You received this message because you are subscribed to a topic in the Google Groups "ROS/Orocos Robot Control Special Interest Group" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ros-sig-robot-control/y0CJ08aVSsM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ros-sig-robot-control+unsubscri...@googlegroups.com.

[Bence Magyar]

I'm not sure the IRC analogy applies here. At the moment we don't know of/when ROS2 will take over

Have a look at the ROS2 roadmap and whether you need those features or the ones from the current ROS releases. Both will he supported for years to come. 

You received this message because you are subscribed to the Google Groups "ROS/Orocos Robot Control Special Interest Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to ros-sig-robot-control+unsub...@googlegroups.com.

[kar...@osrfoundation.org]

Hello,

sorry for being late on this, I've just seen this thread.

As mentioned already, the current implementation of ros2_control is a PoC, strongly inspired by the ROS1 implementation. The goal was to depict features of ROS2 (lifecycle management, multi-node apps) and how they could be used in a controller framework. If you plan on experimenting with ROS2, I encourage you to try it out. However, if you're in a ROS1 environment, ros_control is your choice :) 

To unsubscribe from this group and all its topics, send an email to ros-sig-robot-control+unsub...@googlegroups.com.

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