[cob_scan_unifier] cob_scan_unifier is not merging complete scans in simulation

[benmaidel]

From @ipa-bnm on September 17, 2015 12:5

scan unifier produce gap on the back of cob4-2

@ipa-frm @ipa-mig @fmessmer FYI

Copied from original issue: ipa320/cob_navigation#53

[benmaidel]

From @ipa-mig on September 18, 2015 6:1

hmm, also, there is a z-offset (which will probably not be a problem for our use-cases... but we should change that anyways...

[benmaidel]

the z-offset seems to be ok. Imagine the scanners are mounted in different z positions (eg. our BMW). So transform the scans to the ground is ok, from my opinion. A better solution would be the usage of a point cloud instead a merged scan message.

[benmaidel]

https://github.com/iralabdisco/ira_laser_tools
there you can find a tool with the same purpose as cob_scan_unifier.
Difference:

• no gaps
• dynamic reconfigure
• offers point cloud 2 as well

Perhaps we should stick to this package? We could even copy the code from this package to our repo, because it has no license at all.

[destogl]

@ipa-bnm yesterday I ported cob_scan_unifier to tf2 (s. #257), are you proposing to use this other repository and let cob_scan_unifier to die out? This is important to me to know in which direction to go.

Regarding this not merging scans, I also saw it with real scans but can not really say where is the problem.

[benmaidel]

@destogl for now we will stick to cob_scan_unifier. For the moment there are no plans to switch to another solution. But if we continue to have problems with cob_scan_unifier , it could be that we change or reimplement it. I will also have a look at your PR

[mgruhler]

Also, check the angle increment, see comment

[fmessmer]

@ipa-mig, is @mig-em working on this after we merged #324?

[mgruhler]

he should be :-)

[ghost]

I can't reproduce this bug in the current scan_unifier. I tested by rotating in a corner of ipa-apartment and farer away to increase the effect, but no gaps showed up.

[fmessmer]

@ipa-bnm @ipa-mig
Have you had a chance to test the latest version of cob_scan_unifier (using synchronizer)?
So has this issue become obsolete?

[HannesBachter]

this occured today:

is this related to the issue? (latest commit is from April 5th 2017)