robot_reach_study_node crash if there are too much points in point cloud

[vicltz]

I noticed that the robot_reach_study_node crashes if there are too much points in the point cloud. I thought it was a problem with my robot, but I was able to reproduce it with the demo.

The crash actually comes from the system that kills the process, because it consumes too much RAM (exit code -9)

The crash can occur either during the reach_study (+6000 points in the point cloud , my setup has 16GB of RAM) , or during the optimization loop (+4000 points in the point cloud).

I am not sure if it might be because of the design of the reach algorithm by itself or because of a memory leak your might have.

Note: probably related to the boost object that stores the results, because I was able to get this type of crash (too much RAM used) by loading reach.db.xml with a lot of markers inside (since the reach_study and optimization was already done).