Jack000's profile - activity

in the camera calibration tutorial they use a very large board with a checkerboard pattern. Is a human-sized checkerboard really necessary? It seems to me a smaller checkerboard should be equivalent as long as the camera focus was correct.

Can I use a checkerboard on a letter-sized sheet of paper with good results?

I'm having some problems tuning the local planner on my robot, so I'm trying to simulate in rviz without sensor data.

how do I populate the local costmap with obstacles from the global costmap?

my params look like this right now:

costmap_common.yaml:

obstacle_range: 2.5
raytrace_range: 3.0
footprint: [[ 0.15,  0.15], [-0.15,  0.15], [-0.15, -0.15], [ 0.15, -0.15]]
inflation_radius: 0.30

global_costmap.yaml:

global_costmap:
  global_frame: map
  robot_base_frame: base_link
  update_frequency: 5
  publish_frequency: 1
  static_map: true

local_costmap.yaml:

local_costmap:
  global_frame: odom
  robot_base_frame: base_link
  inflation_radius: 0.3
  update_frequency: 5
  publish_frequency: 2
  static_map: false
  rolling_window: true
  width: 3
  height: 3
  resolution: 0.03
  plugins:
    - {name: static_layer,            type: "costmap_2d::StaticLayer"}
    - {name: inflation_layer,     type: "costmap_2d::InflationLayer"}

but the local costmap is empty so the robot tries to go through obstacles

I think it was just due to dropped frames. Kinectv2 has problems with certain usb hardware apparently and it starts to drop frames after about 30 seconds on my laptop..

I'm building a simple differential drive robot with the kinect+fake laser setup from the tutorials, and now I'm trying to do some navigation with move_base.

my transforms look like this, except instead of IMU I'm combining wheel encoder odometry and visual odometry from RTAB-map.

when visual odometry is used alone, it's mostly fine until the robot gets close to a wall or something. I'm trying to use robot_localization to combine the two odometries and feed it into RTAB-map as /odom_filtered.

en<node pkg="robot_localization" type="ekf_localization_node" name="ekf_localization" clear_params="true">
 <param name="frequency" value="20"/>
 <param name="sensor_timeout" value="1.5"/>
 <param name="two_d_mode" value="true"/>
 <param name="map_frame" value="map"/>
 <param name="odom_frame" value="odom"/>
 <param name="base_link_frame" value="base_link"/>
 <param name="world_frame" value="odom"/>
 <param name="transform_time_offset" value="0.0"/>

 <param name="odom0" value="/fake_odom"/>

 <rosparam param="odom0_config">[false, false, false,
                              false, false, false,
                              true, true, false,
                              false, false, true,
                              false, false, false]</rosparam>
 <param name="odom0_differential" value="true"/>

 <param name="odom1" value="/rtabmap/visual_odom"/>

 <rosparam param="odom1_config">[true, true, false,
                              false, false, true,
                              false, false, false,
                              false, false, false,
                              false, false, false]</rosparam>
 <param name="odom1_differential" value="false"/>

</node>

With robot_localization, after a while the map looks like this:

the 3d and 2d map becomes chaotic and shifts around each frame.

my view_frames looks like this:

any idea what I'm doing wrong? The visual odometry is only published at around 1hz, could that be a problem?

ah, did not know about rqt_graph. I had put the kinect bridge and rtabmap stuff in one launch file, which apparently produced some kind of conflict. I've got grid_map working now. thanks for the help

I'm running this launch file for the kinect2_bridge: https://github.com/introlab/rtabmap_r...

but I don't see any configs that should conflict with generating a 2d map

I have RGB-D images and /scan working, but my expectation was that RGB-D + /scan = grid_map. rtabmap node creates the topic /rtabmap/grid_map but nothing is published to it.

I guess it's a bit confusing because there's no tutorial or demo showing how grid_map works, or even how to enable it.

how do I know which topics are necessary for grid_map? I thought only /scan is required for the grid map

I'm trying to get /rtabmap/grid_map working. I'm using the kinect + fake 2d laserscan method in the tutorial, and there is data being published to /scan

When I rostopic echo /rtabmap/grid_map, nothing is displayed.

I expected that if the subscribe_laserScan argument is on, rtabmap will output a 2d occupancy grid map under /rtabmap/grid_map. Is that incorrect? What other settings could affect grid_map?

just getting started with ROS. I've got RTAB-map working with a kinectv2, but I'm seeing loss of tracking when there aren't enough feature points in the camera view.

I was wondering if adding odometry input would alleviate this problem in the actual robot? Would a laser scanner improve accuracy as well?

I'm contemplating buying a laser scanner as per: http://wiki.ros.org/rtabmap_ros/Tutor...

but I'm not certain how the laser scan interacts with RTAB-map. It says that the laser scan is used to construct a 2d occupancy grid, but that can already be done without RTAB-map. Is the 2d laser scan data used to improve performance in the 3d SLAM at all?

ok, I'll try that. The diagram on the tutorial had both at 3.0 so I thought that was the best way.

I've also tried powering the lidar and motor from 2.8->3.3 volts on a bench supply. At extreme ranges I stop getting data, the debug output is from 3.0v. I've checked to make sure current isn't being limited, it draws about 100mA

I don't have the hardware connected now but /rpms is also empty I think. I'll check it when I hook it up again.

yeah, the /scan topic only appears after running the driver. I've tried both versions of the firmware too.

I'm following these tutorials on using the xv11 lidar with ROS:

http://wiki.ros.org/xv_11_laser_drive...

http://wiki.ros.org/xv_11_laser_drive...

looking at the raw input on /dev/ttyUSB0

2016-02-02 18:41:32.959795: /dev/ttyUSB0
53 43 73 20 6F 72 20 42 52 45 41 4B 20 74 6F 20 SCs or BREAK to 
61 62 6F 72 74 2E 3A 29 0A 0D 50 69 63 63 6F 6C abort.:)..Piccol
6F 20 4C 61 73 65 72 20 44 69 73 74 61 6E 63 65 o Laser Distance
20 53 63 61 6E 6E 65 72 0D 43 6F 70 79 72 69 67  Scanner.Copyrig
68 74 28 63 29 20 32 30 30 39 2D 32 30 31 20 4E ht(c) 2009-201 N
65 61 74 6F 20 52 6F 2E 0A 0D 41 6C 6C 20 52 69 eato Ro...All Ri
67 68 74 73 20 52 65 73 65 72 76 65 64 0A 0D 0A ghts Reserved...
0D 4C 6F 61 64 65 72 09 56 32 2E 35 2E 31 34 30 .Loader.V2.5.140
31 30 0A 0D 43 50 55 09 46 32 38 30 32 78 2F 63 10..CPU.F2802x/c
30 30 31 0A 0D 53 65 72 69 61 6C 09 57 54 44 33 001..Serial.WTD3
34 34 31 32 41 41 2D 31 35 32 36 35 30 0A 0D 4C 4412AA-152650..L
61 73 74 43 61 6C 09 5B 35 33 37 31 32 36 43 5D astCal.[537126C]
0A 0D 52 75 6E 74 69 6D 65 09 56 32 2E 36 2E 31 ..Runtime.V2.6.1
35 32 35 0A 0D 23 53 70 69 6E 2E 2E 2E 33 20 45 525..#Spin...3 E
53 43 73 20 6F 72 20 42 52 45 41 20 74 6F 20 61 SCs or BREA to a
62 6F 72 74 2E 3A 29 0A 0D 50 69 63 63 6F 6C 6F bort.:)..Piccolo
20 4C 61 73 65 72 20 44 69 73 74 61 63 65 20 53  Laser Distace S
63 61 6E 6E 65 72 0D 43 6F 70 79 72 69 67 68 74 canner.Copyright
20 63 29 20 32 30 30 39 2D 32 30 31 20 4E 65 61  c) 2009-201 Nea
74 6F 20 52 6F 6F 74 69 63 73 2C 20 49 6E 63 2E to Rootics, Inc.

it looks like I'm getting data, but

rostopic hz /scan

is empty, as is

rostopic echo /scan

when I

rostopic list

I see

/rosout
/rosout_agg
/rpms
/scan

and it's the same for both versions of the firmware

I'm thinking that the driver might have problems with ROS Jade? I'm not sure what else could be going wrong.

Has anyone else got this working with ROS Jade?

are there any other drivers I could try?