SLAM

The SLAM (Simultaneous Localisation and Mapping) team develops the algorithms that allow our autonomous vehicle to understand where it is and build a map of the track in real time. By fusing sensor data from LiDAR, IMU, and odometry, we provide a reliable localisation backbone that the rest of the autonomy stack depends on.

---

Focus Areas

• LiDAR-based mapping: Processing point clouds to build accurate 2D/3D maps of the cone layout.
• State estimation: Fusing IMU, wheel odometry, and visual data for robust pose estimation.
• Loop closure: Detecting revisited areas to correct accumulated drift over multiple laps.
• ROS 2 integration: Publishing real-time pose and map data for downstream planning modules.

---

Practices

• Simulation-first development: Validating algorithms in simulation before on-car deployment.
• Benchmarking: Comparing localisation accuracy across datasets and conditions.
• Cross-team collaboration: Working closely with Perception and APC to ensure consistent coordinate frames and data flow.