vhwanger's profile - activity

I'd like to use FCL as a collision checker in my own planner that's currently not plugged into Moveit. Is there any documentation or code out there that can help me integrate FCL into my planner? I'm going to be testing on the PR2, so I'd imagine people have done this before Moveit, but I can't find any information.

How do I change the format of ROS_INFO/ROS_DEBUG, etc so that it prints out the filename and line number that it was called from?

I'm trying to view some simple int variables in GDB, but they're all being marked as "value optimized out". At the top of my CMakeLists.txt file, I have

set(ROS_BUILD_TYPE Debug)

Do I need to set something else?

I'm picking out a laptop that will be exclusively used for ROS development. Does anyone have any recommendations for a specific model? I've got my eye on a Thinkpad T530, but I've heard of problems with Nvidia Optimus. I saw that a new driver has been released this month - how well does that work?

Edit: I'm going to be doing a lot of RViz/PCL stuff, so I need some horsepower

Edit: Ah hah, this turned out to be a package problem, not a rospack problem. Sorry!

I have a package in an overlay. This same package exists in the /opt stacks. My ROS_PACKAGE_PATH has the overlay before the /opt stacks.

/home/vchwang/ros_workspace/sandbox:/opt/ros/electric/stacks:/opt/ros/electric/ros

However, roscd still points to the one in /opt. As a result, rosmake refuses to build the package in my overlay. How do I change this?

Some more info:

Even after running rospack profile, my rospack_cache claims the package sbpl is in /opt/ros/electric/stacks/arm_navigation/sbpl. If I go into ~/ros_workspace/sandbox/sbpl and rosmake, I get:

[ rosmake ] No package selected and the current directory is not the correct path for package 'sbpl'.

I'm trying to replay a saved trajectory by using the joint_states positions and JointTrajectoryAction. However, for some of the joints (upper arm roll), the /joint_state accumulates all the movement, so its position is some arbitrarily large number (let's say 600). However, JointTrajectoryAction seems to work on a different frame, because if I issue a goal to position 600, it ends up spinning indefinitely despite the /joint_state claiming it's very close to 600. How do the joint_positions differ?

I've got two Hokuyo laser scanners, both of which work individually when I run the hokuyo_node and change to the appropriate port. However, I'd like to have them both running at the same time, publishing to two different topics. How do I go about this?

I'm trying to hook up costmap2d with octomap's projected_map topic, but I can't seem to figure out how to get things working. All I really need is for costmap2d to inflate and crop a static map and return it as an OccupancyGrid. I've got the following config:

global_frame: /map
robot_base_frame: base_link
update_frequency: 1.0
publish_frequency: 1.0
map_topic: /projected_map

#set if you want the voxel map published
publish_voxel_map: true

#set to true if you want to initialize the costmap from a static map
static_map: true

#START VOXEL STUFF
map_type: costmap
origin_z: 0.0
z_resolution: 0.2
z_voxels: 10
unknown_threshold: 10
mark_threshold: 0
#END VOXEL STUFF

transform_tolerance: 0.3
obstacle_range: 2.5
max_obstacle_height: 2.0
raytrace_range: 3.0
footprint: [[-0.325, -0.325], [-0.325, 0.325], [0.325, 0.325], [0.46, 0.0], [0.325, -0.325]]
#robot_radius: 0.46
footprint_padding: 0.01
inflation_radius: 0.55
cost_scaling_factor: 10.0
lethal_cost_threshold: 100
observation_sources: base_scan
base_scan: {data_type: LaserScan, expected_update_rate: 0.4,
  observation_persistence: 0.0, marking: true, clearing: true, max_obstacle_height: 0.4, min_obstacle_height: 0.08}

When I start up both nodes (octomap server and costmap2d), I don't get an active gridcell topic in rviz. Also, I'm not running this with Gazebo, if that matters. Any ideas?

I'd like to use the projected map generated from octomap for some 2d stuff, but when I look at the map returned as a nav_msgs/OccupancyGrid, it's of size 1001 x 1001 (when my map is half that). Is there some way to crop the map to a defined size?

Unfortunately, I also double as the network admin. Might this have to do with us not using the network boot for the base station? We filed a ticket, but it has gone unattended to.