Project

General

Profile

Actions

Bug #336

closed

Emulation of bidirectional rotors

Added by Youssef Aboudorra 8 months ago. Updated 8 months ago.

Status:
Closed
Priority:
Normal
Assignee:
-

Description

Hello Anthony,

Not a bug report per se. However, I am working on a controller that requires rotor velocities in both directions, and I was wondering if something could be done to let mrsim-gazebo emulate this.

So far I use rotorcraft-genom3, and seems to me it accepts the negative rotor_input without issues, but it's not clear to me if ignores the sign of direction when it communicates with the pty of mrsim-genom3, or mrsim-genom3 is agnostic to the the sign.

I would appreciate your insight on this.

Regards,
Youssef

Actions #1

Updated by Antonio Franchi 8 months ago

just a couple of additions to Youssef message:
  • for a project we need to deliver a work in which we use bidirectional propellers in which we need to do simulations, simulations have to be done no later than this month (11/21)
  • if it is not possible in mrsim then we need to use another simulator
  • so it is very important for the project to know if this is possible/available in mrsim otherwise we need quickly to migrate to another simulator or modify the existing one in some other way
Actions #2

Updated by Anthony Mallet 8 months ago

Right, there are several issues that curently prevent from inverting
the spipnning direction. The sign of the target velocity is ignored in
several places.

Thanks for the report, I will check if this can be fixed easily.

Actions #3

Updated by Anthony Mallet 8 months ago

  • Status changed from New to Feedback

Could you please test the last commits and report if it fixed the
issue for you?

You need libmrsim:9422b9392 and mrsim-gazebo:e5432d5c4

Actions #4

Updated by Youssef Aboudorra 8 months ago

Hello Anthony,

Thanks, for your fast response.

Now the spinning velocity can be inverted along with the corresponding thrust, however this was only possible when I send the set points using set_velocity() of rotorcraft.

If I use the servo() function to rely on the rotor_input port connected to the controller component, and a switch in the spinning is commanded from the controller: all rotors stop spinning. So I think the change of spin triggers something in servo() to stop the rotors.

Appreciate your support

Actions #5

Updated by Anthony Mallet 8 months ago

Thanks for testing.

When rotors stop, could you check if the servo() activity reports a
"e_rotor_failure" exception?

Since while reverting, the rotors temporarily pass through a "not
spinning" state, this might be the explanation. I'll try to find a
proper fix.

Actions #6

Updated by Anthony Mallet 8 months ago

In the meatime, I attempted a fix in rotorcraft-genom3.

You might want to try the last update, at least
rotorcraft-genom3:39d9a11615. For me it seems to work (although
I don't have any suitable controller for in-depth testing).

Actions #7

Updated by Youssef Aboudorra 8 months ago

Hello,

Thanks Anthony!
Now it works with the last fix. I will keep testing and if I find something strange, will report it.

P.S. I have another point regarding the max spinning velocity, but I will post in another issue to conclude this one.

Thank a lot for your support

Regards,
Youssef

Actions #8

Updated by Anthony Mallet 8 months ago

  • Status changed from Feedback to Closed

Good! I'm closing the issue, then.

Actions

Also available in: Atom PDF