LDRMS scans published in frame that does not exist in tf tree

[miguelriemoliveira]

@danifpdra can you help me? You use the sick ld rms right? Do you subscribe to the scan0, 1, etc messages?

[miguelriemoliveira]

[danifpdra]

@miguelriemoliveira I included the launch file for the lSICK LDMRS and the Novatel GPS + IMU in the altas-core in the car, and it is all lauched with the drivers.launch file (I didn't commit to github tho)

To obtain sensor data, topics /ld_rms/scan0 /ld_rms/scan1 /ld_rms/scan2 /ld_rms/scan3 must be subscribed, but to group all the data from the sensors, it is necessary to create a point cloud assembler like this:

    <node type="laser_scan_assembler" pkg="laser_assembler" name="my_assembler0">
        <remap from="scan" to="/ld_rms/scan0"/>
        <remap from="assemble_scans2" to="assemble_scans0"/>
        <param name="max_scans" type="int" value="100" />
        <param name="fixed_frame" type="string" value="map" />
    </node>
    <node type="laser_scan_assembler" pkg="laser_assembler" name="my_assembler1">
        <remap from="scan" to="/ld_rms/scan1"/>
        <remap from="assemble_scans2" to="assemble_scans1"/>
        <param name="max_scans" type="int" value="100" />
        <param name="fixed_frame" type="string" value="map" />
    </node>
    <node type="laser_scan_assembler" pkg="laser_assembler" name="my_assembler2">
        <remap from="scan" to="/ld_rms/scan2"/>
        <remap from="assemble_scans2" to="assemble_scans2_"/>
        <param name="max_scans" type="int" value="100" />
        <param name="fixed_frame" type="string" value="map" />
    </node>
    <node type="laser_scan_assembler" pkg="laser_assembler" name="my_assembler3">
        <remap from="scan" to="/ld_rms/scan3"/>
        <remap from="assemble_scans2" to="assemble_scans3"/>
        <param name="max_scans" type="int" value="100" />
        <param name="fixed_frame" type="string" value="map" />
    </node> 

and maybe see this file https://github.com/lardemua/road_detection/blob/master/negative_obstacles/src/cloud_assembler.cpp

[miguelriemoliveira]

Hi @danifpdra ,

Thank you for the help. One thing I don't understand ... how can you assemble without knowing where the frame /ld_rms/scan1 (for example) is?

Can you post the pdf that results from

rosrun tf view_frames

Do you have the link ldrms0 (check my image, that's the frame_id of the messages) in your tf tree?

[danifpdra]

frames.pdf

@miguelriemoliveira , I do have the ldrms frames. Is the device_frame_publisher node from the free_space_detection package running?

[miguelriemoliveira]

Not sure. Let me check your drivers.launch ...

[miguelriemoliveira]

Hum, I don't think there is any device_frame_publisher in the atlas core repo. Where do you have it?

[danifpdra]

The drivers.launch only launches the sensor... I launch that node on my own launch file.

[danifpdra]

The free space detection package is in the lardemua github, it was created in previous years

[miguelriemoliveira]

OK, can you tell me where it is?

Also, it is clearly something that belongs to the scope of the drivers. It should be added to the drivers.launch. Can you add it to drivers.launch and remove it from your launch file?

[danifpdra]

Here: https://github.com/lardemua/free_space_detection

Yes, I'll do it in the atlascar computer since the github is not updated

[miguelriemoliveira]

It should be. If you can call me when you go and I will try to go with you.

Thanks, Miguel

• Miguel Riem de Oliveira email: mriem@ua.pt | m.riem.oliveira@gmail.com Dep. de Engenharia Mecânica, Universidade de Aveiro Campus Universitário de Santiago, 3810-193 Portugal*

On Tue, 28 May 2019 at 00:39, Daniela Rato notifications@github.com wrote:

Here: https://github.com/lardemua/free_space_detection

Yes, I'll do it in the atlascar computer since the github is not updated

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/lardemua/atlas-core/issues/11?email_source=notifications&email_token=ACWTHVVVH4ELRR47IHVIFJ3PXRWLDA5CNFSM4HP6RJ42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWKUFAQ#issuecomment-496321154, or mute the thread https://github.com/notifications/unsubscribe-auth/ACWTHVQY3VICJIJB45OP3LDPXRWLDANCNFSM4HP6RJ4Q .

[miguelriemoliveira]

Updated in 07d32af302aee6eb57e0b033f5f57e0f1686b8cc

[miguelriemoliveira]

So investigating this a bit more I noticed we are using a driver which is not the most common ...

@bernardomig you have any idea why?

It should be this one

https://github.com/SICKAG/sick_ldmrs_laser

since it is the one that appears officially in the ros webpages ... the problem we detected is because the "non-oficial" driver publishes the laser scans using a frame_id which does not exist in the tf tree.

The oficial driver does not publis the 2D scans separately, it only publishes a cloud.

I think this one would be better to use. @danifpdra do you need the separate scans? Is the point cloud enough for you?

I propose we change to this other driver, tell me what you think ...

[miguelriemoliveira]

here is the drivers.launch ...

    <!--    TODO can we replace this by the oficial driver?-->
    <node pkg="sick_ldmrs" type="sickldmrs.py" name="ld_rms" required="true" output="screen" if="$(arg ld_mrs)">
        <param name="host" value="192.168.0.244"/>
        <param name="port" value="12002"/>
    </node>
    <!--    TODO: Daniela chekc if this works for you?-->
    <!--    <group ns="frontal_laser">-->
    <!--        <include file="$(find atlas2_bringup)/launch/sick_ldmrs_node.launch">-->
    <!--        </include>-->
    <!--    </group>-->