Problem implementing new state space

[ompl]

Problem implementing new state space

[mamoll]

Original comment by Ioan Sucan (Bitbucket: isucan, GitHub: isucan).

---

The issue is that the R3 component in SE3 needs some bounds to be specified when you call setup (automatically called when planning). Based on those bounds some computation is performed to determine the step size for checking collisions. If bounds are not set, the computation cannot be performed. I think all you have to do is call setBounds() on the SE3 component (which forwards the call to the R3 component)

[mamoll]

Original comment by neckutrek (Bitbucket: neckutrek, GitHub: neckutrek).

---

I added this code:

#!c++

void LinkQuad_Setup::setDefaultBounds()
{
  std::cerr << "setDefaultBounds()\n";
  ob::RealVectorBounds se3bounds(3);
  se3bounds.setLow(-10.0);
  se3bounds.setHigh(10.0);
  this->getStateSpace()->as<LinkQuad_StateSpace>()->as<ob::SE3StateSpace>(0)->setBounds(se3bounds);
}

And now I get this output instead:

setDefaultBounds()
terminate called after throwing an instance of 'ompl::Exception'
what(): The longest valid segment for state space RealVectorSpace10 must be positive

So it went from "RealVectorSpace8" to "RealVectorSpace10"...

Do I have to set bounds for the 1D RealVectorStateSpace at index 1 of my CompoundStateSpace as well? The code for that special state space is avaiable in my previous post!

Thanks!

[mamoll]

Original comment by Ioan Sucan (Bitbucket: isucan, GitHub: isucan).

---

Hey, you have to set those bounds as well, yes. You should also call the printSettings() method on your state space. It should show you information about what bounds are set.
If you like graphviz, you might like the Diagram() function as well.
This is becoming more of a question rather than an issue, so I will close the issue now. Please reopen if you find a problem or e-mail on the mailing list if you have further questions.