Fast RRT*
Enhanced the RRT* sampling-based motion planner by replacing its linear-scan nearest-neighbour and near-neighbour queries with a Kd-tree index, reducing per-iteration cost from O(n) to O(log n). Achieved substantially faster convergence to high-quality paths on cluttered 2D planning benchmarks, enabling practical use in larger maps where vanilla RRT* became prohibitively slow.
Dixant Mittal
My research interests include reinforcement learning, planning & search, large language models, and decision-making under uncertainty.