Victor22's profile - activity

Hi all: I am using octomap in my project. I want to clear a region inside given bounding box in the 3D map constructed using Octomap. I know that there exists a service ("clear_bbx") in Octomap to clear a region inside bounding box, however this service only clears the occupied voxels but not the unknown region within bounding box.

I know that the unknown voxels are not initialized and therefore it (clearing of unknown) requires to iterate through them and thereafter further divide an unknown voxel (till it fits into bounding box or no need if it is already in bounding box) and set its occupancy value of FREE.

Need help on this Please. Thanks!

Hello: I just installed ROS Fuerte in Ubuntu 12.04 and doing simulations in Gazebo. I spawned my robot model and kitchen environment in Gazebo. Whenever I stop executing the code it gives me an error, which is as follows:-

terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::lock_error> >' what(): boost::lock_error

As such I don't have any problem with this error, my code runs smoothly, but whenever I stop my code (ctrl+C), then it gives me this error.

Apart from that, I also get these warnings in Gazebo when I spawn my robot model or environment model, Also if I spawn more than 2 models (say tables) then my Gazebo used to crash:

Error [ColladaLoader.cc:1392] No Opaque set Warning [Visual.cc:689] Unable to get Material[Gazebo/LightWood] for Geometry[iros_table_model::table_top_link::table_top_link_geom_visual. Object will appear white

I just switched from Electric, there I did not have above mentioned errors with the same files (robot model, environment model and my code).

I checked my boost library version:

$dpkg -s libboost-dev | grep 'Version'
Version: 1.48.0.2

I do not know whether I should put this question here or on Gazebo forum. Please guide me.

Hello: I am looking for end-effector pose constrained planning for the manipulator. I know that such functionalities have been provided in "MoveIt!" where you set your kinematics constraints and then give command to plan a path.

I believe that this type of end-effector constrained planning (in MoveIt!) is done in end-effector cartesian degree of freedom using Inverse Kinematics.

Can anyone please point me to: where this task-constrained code sits in OMPL (as MoveIt uses OMPL), or any pointer to the research paper of this implementation in MoveIt!

I want to understand how it is done and what are the steps involved in ROS implementation.

Thanks.