docs(autoware_lidar_frnet): add UML diagram for processing flow

[kyoichi-sugahara]

Description

I have added a processing flow diagram to visualize the logic flow. This addition is intended to facilitate future discussions regarding code improvements, such as multi-threading and batch processing optimization.

Note:
The diagram is designed for a high-level understanding and does not aim to be strictly comprehensive or cover every single processing step; therefore, some details may be omitted.

I am open to adjusting the granularity of the processes described in the flow. If there are specific policies or preferences regarding how detailed these steps should be, please provide feedback in the comments.

Related links

Parent Issue:

• Link

How was this PR tested?

Notes for reviewers

None.

Interface changes

None.

Effects on system behavior

None.

[github-actions]

Thank you for contributing to the Autoware project!

🚧 If your pull request is in progress, switch it to draft mode.

Please ensure:

• You've checked our contribution guidelines.
• Your PR follows our pull request guidelines.
• All required CI checks pass before marking the PR ready for review.

[github-actions]

Documentation URL: https://autowarefoundation.github.io/autoware_universe/pr-11660/
Modified URLs:

• https://autowarefoundation.github.io/autoware_universe/pr-11660/perception/autoware_lidar_frnet/

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 17.78%. Comparing base (d8dec0f) to head (b204a36).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #11660      +/-   ##
==========================================
- Coverage   17.79%   17.78%   -0.02%     
==========================================
  Files        1747     1748       +1     
  Lines      122053   122099      +46     
  Branches    42794    42813      +19     
==========================================
- Hits        21723    21714       -9     
- Misses      82134    82214      +80     
+ Partials    18196    18171      -25     

Flag	Coverage Δ		*Carryforward flag
daily	19.91% <ø> (+<0.01%)	⬆️	Carriedforward from d8dec0f
daily-cuda	17.80% <ø> (+<0.01%)	⬆️	Carriedforward from d8dec0f
differential-cuda	0.00% <ø> (?)
total-cuda	17.80% <ø> (+<0.01%)	⬆️	Carriedforward from d8dec0f

*This pull request uses carry forward flags. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[amadeuszsz]

@kyoichi-sugahara
Thank you for your PR!

We definitely need such a inner-algorithm description. However, since proposed diagram is not C++ UML architecture design (not enough details), in some places it seems to me too detailed which increases maintenance cost.

For example, you could just use a single block "Copy outputs from device to host" instead of pointing each tensor separately.

On the other hand, diagram is not detailed enough, or let's say - too ambiguous. The greatness of such a diagrams is that most of users can take a look and understand the whole pipeline. Let me start from beginning of the diagram:

• Input validation does not show what happens if input is invalid.
• I don't what "against profile" means. What is the profile? (I know, but any user even with technical background will get confused). What happens if input is invalid? Wouldn't be better to write it as "Check the input point cloud’s point count against the min/max thresholds".
• What yaw and pitch angles refer to? You could write that it is about sensor's specification, but having in mind previously described detail level, "Project 3D points to 2D frustum" is enough.
• And so on.

Please consider to update this diagram.

[kyoichi-sugahara]

@amadeuszsz
Thank you so much for your comment!
I agree with your feedback!

Can I try updating diagram based on your feedback? Or, if you don't think this diagram is strictly necessary, I'm fine with closing this PR 😅 because the main purpose of this PR was to clarify my understanding, and I'm already satisfied.