Localization Zoo

Localization research, mapped

Choose, compare, and reproduce localization methods.

C++ localization paper implementations, derived variants, and compact baselines behind a shared benchmark surface.

Run the 3-minute demo Explore Methods View on GitHub

No build or dataset required. The first image download may take longer.

Methods -
Families -
CI Fixture 3 frames
Latest benchmark trajectory overlay
Latest benchmark snapshot Loading...

Current Benchmark Results

Loading benchmark...

Accuracy vs Throughput

Benchmark Map

ATE uses log scale; filter by benchmark

Leaderboard

Odometry Results by Benchmark

Rankings are computed only within the same benchmark group. Smoke checks and GT-seeded references are shown separately.

Method Benchmark Status Policy ATE [m] RPE [%/100 m] Coverage Total elapsed [ms] FPS Note

Method Explorer

Find the right starting point

Family
Signal

Starter Stack

Quick Smoke

    3-minute demo

    Run a real comparison, no setup required

    No usage data is sent

    Choose your terminal, paste one command, then open the generated report. Docker runs the prebuilt quick profile against bundled fixtures.

    1Run

    Docker downloads the image and runs the quick comparison.

    2Confirm

    The terminal ends with == Done ==.

    3Continue

    Open zoo-demo/report.html and choose your next run.

    Prefer a native build? 
    git clone https://github.com/rsasaki0109/localization_zoo.git
cd localization_zoo
sudo apt install libeigen3-dev libpcl-dev libopencv-dev libceres-dev libgtest-dev
bash evaluation/scripts/demo_localization_zoo.sh

    This page keeps anonymous button counts and first/last visit days in this browser only. It never stores searches or command contents and sends no analytics.