feat(map_based_prediction): apply the custom find-nearest function

[ktro2828]

Description

Because the builtin lanelet2::findNearest(...) function doesn't consider the 3D spaces, it tend to take a longer time and consume higher computational costs in the case that lanelet has 3D dimensional overlaps.

In autowarefoundation/autoware_core#839, the custom fined nearest function is added.

Key changes:

Lanelet type updates for const-correctness:

• Changed function signatures and variable types from lanelet::Lanelet to lanelet::ConstLanelet in both header (utils.hpp) and implementation (utils.cpp) files, including in calculateLocalLikelihood and checkCloseLaneletCondition. [1] [2] [3]

Nearest lanelet search improvements:

• Replaced the use of lanelet::geometry::findNearest with experimental::lanelet2_utils::find_nearest for finding surrounding lanelets, which now returns ConstLanelet pairs.

Const removal where mutable lanelets are needed:

• Used experimental::lanelet2_utils::remove_const to convert ConstLanelet to a mutable Lanelet where necessary for downstream processing. [1] [2]

Include updates:

• Added the autoware/lanelet2_utils/nn_search.hpp header to provide the new nearest neighbor search functionality.

Related links

Private Links:

• TIER IV INTERNAL

How was this PR tested?

Checkout autoware_core to ktro2828:feat/lanelet2_utils/find_nearest and build:

cd src/core/autoware_core
gh pr checkout 839
cd ../../../
colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release --packages-up-to autoware_map_based_prediction

The following view shows the prediction visualization comparison (left: before, right: after).
Although the previous results are flicking due to high latency, the updated results are stable.

Screencast.from.2026.02.18.18.09.39.webm

The following figure shows the performance profile comparison:

Notes for reviewers

THIS PR MUST BE MERGED AFTER autowarefoundation/autoware_core#839 HAS DONE!!

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.

[YoshiRi]

LGTM but could you add some comments?

[YoshiRi]

LGTM

[codecov]

Codecov Report

❌ Patch coverage is 0% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 18.61%. Comparing base (ae88de5) to head (a56ffd7).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
...eption/autoware_map_based_prediction/src/utils.cpp	0.00%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #12128      +/-   ##
==========================================
+ Coverage   18.11%   18.61%   +0.50%     
==========================================
  Files        1841     1872      +31     
  Lines      126643   127654    +1011     
  Branches    44419    46807    +2388     
==========================================
+ Hits        22941    23764     +823     
- Misses      84592    85340     +748     
+ Partials    19110    18550     -560     

Flag	Coverage Δ		*Carryforward flag
daily	20.61% <ø> (+0.01%)	⬆️	Carriedforward from a2ee3ef
daily-cuda	18.14% <ø> (+0.03%)	⬆️	Carriedforward from a2ee3ef
differential	2.33% <0.00%> (?)
differential-cuda	3.19% <0.00%> (?)
total-cuda	18.11% <ø> (+<0.01%)	⬆️	Carriedforward from a2ee3ef

*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.