alansrobotlab / inmoov_ros Goto Github PK

Hello Alan,
I have been using your URDF as the basis of my project, and forked from your repository a couple of weeks ago to create my own. However, I was wondering what license you were using for your repository; We here at the University of Cincinnati were hoping to use the BSD-3 license.

Do you need to make any modifications(add sensors) to Inmoov to get it to work with your ROS program?

Trainer is currently hardcoded to support (12) servos across (3) teensy/arduino in a specific configuration.

While the motorcommand messages shouldn't be a problem, as there's another enhancement to abstract the message queue, not quite sure how to handle the motorparameter, as they're tied to a service queue, and a specific arduino_serial instance.

It may make sense to keep the bus/servo drill down more or less standard(left arm, right arm, torso), but it needs to support alternative arrangements, at the very least the standard inmoov config of arduino mega + uno.

Implement common queues for all three topics, such that the arduino implementation is hidden.

Queues will be configured (undecided how) based on how the specific bot is configured.

To make things easier for newcomers, perhaps the default configuration will be the inmoov arduino mega + uno default.

Also create a simple UI to enable and disable queue forwarding to both urdf and the physical bot.

Hi, I tried to run the moveit-display.launch file, but it appears there are no moveit files in the repo. Is there a way to launch move group to control the arms here?

Hi Alan,

We have built the InMoov Robot and had a successful run with the MyRobotLab implementation.

Now we are planning to run the InMoov on a ROS stack.

Can you please advise on any hardware modifications needed ( especially Servos part ) to make the InMoov work on ROS.

Adding to my doubt, what is the use of feedback here? I presume that the hardware modifications are needed to get the feedback from servos. Please advise.

PS: We are new to ROS. So not sure how to proceed.

Thanks and Regards,
Gautam

Hi, I'd like to make a telepresence node with 1:1 correspondence. I was trying to go through the inmoov_tools files and I noticed that there are some nodes that might be useful, such as headtracking and look. However, the documentation here seems to be lacking, and I'd rather not pour through the code to see if I can salvage it for my use :)

I'm not expecting a response here (strangely everything inmoov related was most recently updated 2 years ago .. i just found out about it lol), but it would be useful for progeny if there was more documentation.

Revise connection between urdf (rviz,gazebo) model and physical bot, such that the solution can support differing orientations between the axis of rotation between the two.

Must take into account that motion planners like (moveit!) will use the urdf model as the basis for their motion plans.

Look.py uses threading for the random inputs, and to feed position updates at 40 per second. signalling is kinda broken right now, program exit works without hanging, but you can't stop and restart either without crashing the program.

Either set it up to correctly stop and restart the threads, or fix the signals, leverage an enable and an exit correctly.

I figure it out the last problem which I asked:
"When I run this "rosrun inmoov_tools trainer.py", it dosent work, the information is:
Traceback (most recent call last):
File "/home/amiao1017/catkin_ws/src/inmoov_ros-master/inmoov_tools/trainer/trainer.py", line 15, in
from inmoov_msgs.msg import MotorStatus
ImportError: No module named inmoov_msgs.msg"

I think is the 'inmoov_ros/inmoov_tools/trainer/trainer.py ':
15#from inmoov_msgs.msg import MotorStatus"
16#from inmoov_msgs.msg import MotorCommand
17#from inmoov_msgs.srv import MotorParameter
shoud be "inmoov_messages.msg" instead of "inmoov_msgs.msg"

btw, the wiki said "rosrun inmoov_tools set_parameters.py #loads parameters into server"
shoud be "rosrun inmoov_tools setup_parameters.py"

Hello Alan
It's impressive that your do this program. I'm very intersting about that.
And now, I finished the inmoov robot and be familiar with the ROS.
I think that the inmoov_ros is very suit, but there are some difficults when I attemt to use it.

1. Can you offer the microcontroller's type and wiring scheme, or something specific.
2. When I run this "rosrun inmoov_tools trainer.py", it dosent work, the information is:
   Traceback (most recent call last):
   File "/home/amiao1017/catkin_ws/src/inmoov_ros-master/inmoov_tools/trainer/trainer.py", line 15, in
   from inmoov_msgs.msg import MotorStatus
   ImportError: No module named inmoov_msgs.msg

Update Arduino sources to support the standard inmoov configuration (Arduino Mega + Arduino Uno)

• Be flexible enough to support other arduino driven configurations.
• Rename package InMoov_Arduino?
• Support enabling/disabling functions that rely on a/d input
  • wrap in ifdef
  • disable moveto on enable() if no initial position defined

If possible, make it eas(ier) for new people by maybe having the default configuration for the standard inmoov arduino setup?

Undo all of the work to adapt the urdf model to have the same joint orientations as my physical InMoov.
It really complicates the urdf file, and at the end of the day the robot doesn't care.

This would be done in conjunction with some of the other issues to decouple the joint commands between the physical bot and the urdf model in rviz.

The desired end state for this issue is to remove all of the 180 degree rotations for the stl files and joints to simplify the xacro files and urdf model. keep selectors for required mirroring of left vs right features if needed.

Mark all packages with BSD-3 licensing.
(Except for stl meshes, handled in different issue.

Separate out all meshes from inmoov_description into inmoov_meshes.
Licence package as CC-NC-A as per original author's terms