Hi, maybe I found a bug in libgazebo. The situation is the following.
Using MATLAB S-function I impose recursively position and attitude to an object (with an helicopter mesh) with 2 cameras mounted on (like a stereo pair). I observed a strange behaviour during the calls to libgazebo. If I impose an attitude (also constant) different from (0, 0, 0), after a variable number of calls (it depends on the entity of the imposed angles) the cameras seem to fall from the helicopter and give deformed images (and then normally the simulation crashes). Sometimes it happens also that the trimesh attached to the helicopter body stretches and deforms. If the imposed attitude is (0, 0, 0), I can move the cameras around with no problem. I didn't find any explanation to this behaviour. A memory error is unlikely (otherwise I can't explain why with (0, 0, 0) values all goes ok!). Maybe it is a sort of conversion error between euler and quaternions. This can explain why the mesh stretches, if the quaternion is not normalized to1! But why it happens only after 50-60 calls of libgazebo and not immediatly? And wht it doesn't happen when euler angles are (0, 0, 0)? If you need, I can send you pictured of the deformed helicopter and camera sight. And I can also send any information about the C code that calls libgazebo. Thank you Domenico ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Playerstage-gazebo mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/playerstage-gazebo
