Turtlebot 3 burguer (with LIDAR).
Camera Orbecc Astra.
I'm trying to use the compressed topics because with normal topics I don't have enough bandwidth for my Astra camera and LIDAR working both at the same time. So this is the changes I did in my launch:
<arg name="rgb_topic" default="/camera/rgb/image_rect_color/" /> <arg name="depth_topic" default=<b>"/camera/depth/image_raw/" /> Not sure If I have to use camera/depth_registered/image_raw because when I use it, my camera/depth_registered/image_raw/_relay dont publish any menssage. Instead with /camera/depth/image_raw my relay publish normally. <arg name="camera_info_topic" default="/camera/rgb/camera_info" /> <arg name="depth_camera_info_topic" default="$(arg camera_info_topic)" />
<arg name="compressed" default="<b>true"/>
<arg name="rgb_image_transport" default="compressed"/>
<arg name="depth_image_transport" default="<b>compressed"/> One question here, is what is the difference between compressed and compresseDepth, beucase when I use compressedDepth i have the next problem:
when subscribing to compressed topics, the name of the topic should be the normal ones (raw), but image_transport parameter should be set to compressed. For depth image, image_transport should be set to compressedDepth to avoid compression artifacts and that the depth image keeps the right format (16 or 32 bits). Make sure compressed_depth_image_transport plugin is installed on all computers.
When using rtabmap.launch, image_transport value is automatically set to compressed or compressedDepth when compressed argument is set to true (so you don't have to manage this yourself). Relays are also automatically created to make sure the image topics are subscribed only once by the remote computer.
On the computer running the camera, see if the default topics used by rtabmap are actually published:
I installed 'sudo apt-get install ros-kinetic-image-transport-plugins' in my TurtleBot3 which contain my camera Astra.
I'm have a lot of issues, lets explain:
FIRST : using topic camera/depth_registered/image_raw
As you can see here, we Z both topic working fine but I cant use them on RTAB because their Hz is so low, always I'm using a topic with sensor_msg/compressedDepth it reduces a lot my Hz in all my topics.
So I decide to use my topic sensor_msg/compressed intead of compressedDepth... but here continue my problem, something in my transports packages is wrong, but i installed my plugins.
SECOND: Nothing to do with topic camera/depth_registered/image_raw so I use now camera/depth/image_raw
I changed my rtab.launch as you can see here. I'm using camera/depth/image_raw/compressed instead of compressed/Depth becasue it reduces a lot my Hz
So when i finally execute RTAB with this topics I have the next issue:
Finally I can use RTAB without problems with my compressed topics, also my robot can localize itself. Everything is fine unless I want to navigate. So... I show you my problem:
This is my Tf_Tree, obviously is wrong:
So I decide to change my frame ID in my rtab.launch
<arg name="frame_id" default="<b>base_link"/> I change my base_ink to Base_footprint. When I do this the TF tree is correct, like this:
But I have one problem... when my tf tree is correct and I execute my navigation.launch (Because my rtab.launch and navitation.launch are not in the same launch, I will copy my nav.launch later if u can see something wrong)
So I tried to change my robot base frame in costmaps to base_link instead of base_footprint.. Im misunderstanding this, could you explain it to me?.
[ WARN] [1528459815.208587375]: Costmap2DROS transform timeout. Current time: 1528459815.2085, global_pose stamp: 1528459811.5396, tolerance: 0.5000
[ WARN] [1528459815.208638059]: Could not get robot pose, cancelling reconfiguration
This is my nav.launch
<arg name="model" default="$(env TURTLEBOT3_MODEL)" doc="model type [burger, waffle, waffle_pi]"/>
As you can see in this image when my frame ID in Rtab launch is base_link, my tf base_footprint (blue square in photo) is out of my robot. No sense. But if I change my frame ID in rtab.launch to base_footprint instead of base_link my tf base_footprint is correct in my robot but my lasers scan start failing. Like this: