EKF-SLAM from scratch

Robotic Perception class project at Toyohashi University of Technology, in Japan: simultaneous localization and mapping with an extended Kalman filter, in a simulated two-dimensional world.

The notebook walks the full loop: world and robot simulation, motion and measurement noise models, filter initialization, the prediction step, the correction step, and the actual SLAM part — adding landmarks the robot has never seen before. The final trajectory and map are compared with ground truth.

Python, NumPy and Matplotlib only. There is no SLAM library: the filter is implemented from scratch.