Solenoid's profile - activity

For future reference: I ended up using move_group python interface according to the tutorial and defining each step myse

Open gripper using GripperCommand in pick and place pipeline I'm looking to use a 6 DOF robotic arm and an attached 2-fi

I am trying to make a path planner using information from a stereo camera. I am running a simulation in V-REP as a first step.

However I have some strange behavior for which I haven't found any solutions: the point cloud is horizontally flipped and slightly offset from where is should be:

As you can see the stereo images have the test checkerboard in the center, but in the point cloud the checkerboard is slightly off to the left... Has anybody experienced the same issue. The point cloud is also horizontally inverted.

The V-REP scene looks like this:

These are the transforms, for some reason camera_left_optical and camera_right_optical are not connected, but I had to explicitly define them to change the camera orientations...:

The disparity from the stereo_image_proc looks like this:

One can clearly see that it is correct, there is no offset and the scene is in the correct orientation.

Finally this is my node graph if it can help:

I followed the standard robot coordinate system. When I try to turn the robot the point cloud moved with the robot (the wrong way), I expect the elements to stay where they are:

Then the next issue is the /scan topic generated by the pointcloud_to_laserscan node which does strange things as one can see on the map, but that could be related to the point cloud issue.

Did you find an answer to your issue, I am in the same situation and cannot figure this out... Joints are published, but transforms are empty, subscriptions are all correct: robot->joints->robot_state_publisher->tf.