Course project for Robotics.
- Implemented a configurable Monte Carlo Localization simulator in Python with differential drive kinematics, ray-casting distance sensing, and Gaussian likelihood weighting.
- Accelerated ray intersection computation by 10x and reduced rendering time by 50% by replacing Shapely API calls with custom vectorized NumPy cross-product logic.
- Improved tracking robustness during sharp turns with a hybrid Low-Variance resampling strategy and exponential moving average weight smoothing.