Efficient and Interpretable Vehicle Controller with Riemannian Motion Policies

This project involves an autonomous vehicle algorithm inspired by Riemannian Geometry and implemented within Unreal Engine.

Won 2022 Regeneron International Science and Engineering Fair (ISEF) Finalist. Link

 

My contributions: 

1. Designing the algorithm using differential geometry principles: Tailored the autonomous navigation system to leverage the mathematical framework of Riemannian Geometry for improved pathfinding and decision-making.

 

2. Integrating the algorithm with Unreal Engine: Established a robust connection between the algorithm and Unreal Engine, enabling real-time simulation and testing of autonomous vehicle behaviors.

 

3. Capturing interactions within the virtual environment: Implemented mechanisms to record and analyze the vehicle's interactions with its surroundings, facilitating the refinement of its navigational strategies.