Overview
What this challenge is about.
You receive 4 ROS bags from real customer plants, each containing 2D LiDAR scans, wheel odometry, and ground-truth poses (from a motion-capture cell used only for evaluation). Implement a particle-filter SLAM (you may use the gmapping or slam_toolbox reference and modify it) and tune it for the plant environments. Report absolute trajectory error (ATE) per bag, particle-cloud convergence time, and CPU usage on a target embedded SBC. The integration memo lists which firmware changes are needed and the expected accuracy in service.
The Brief
What you'll do, and what you'll demonstrate.
Replace the AprilTag-dependent localization with a 2D-LiDAR particle-filter SLAM that's robust to dirty fiducials.
Earning criteria — what you'll demonstrate
- Apply particle-filter methods to a real state-estimation problem
- Integrate LiDAR + odometry sensor fusion in a robotics stack
- Evaluate SLAM accuracy against motion-capture ground truth
- Plan an embedded integration with realistic CPU/memory constraints
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
AI Engineer
End-to-end SLAM node delivery with embedded CPU constraints is the canonical AI-engineer project in industrial robotics.
This challenge sharpens
- slam
- ros
- state-estimation
Computer Vision Engineer
LiDAR + odometry fusion under real-world noise transfers cleanly into perception/CV engineering roles on autonomous-vehicle teams.
This challenge sharpens
- lidar
- state-estimation
- particle-filter
Machine Learning Engineer
Reproducible evaluation against motion-capture ground truth is the kind of measurement discipline MLEs need on any perception system.
This challenge sharpens
- slam
- python
- particle-filter