Trajectory Planning for Autonomous Valet Parking

This work is in collaboration with the Honda Research Institute, USA. I focused on safe and efficient trajectory planning for Autonomous Valet Parking (AVP) in cluttered environments with dynamic agents. We proposed a time-indexed Hybrid A* planner that accounts for dynamic obstacle predictions, generating collision-free paths in diverse parking scenarios such as perpendicular, angled, and parallel parking. To handle partial observability and evolving spot availability, our framework integrates a probabilistic estimator that predicts future spot occupancy and guides exploratory navigation, balancing goal-directed maneuvers with information-gathering behaviors. Simulations show that this integrated approach improves parking efficiency by 20%, and generates 2x smoother and safer paths compared to existing methods.

Farhad Nawaz, Minjun Sung, Darshan Gadginmath, Jovin D’ sa, Sangjae Bae, David Isele, Nadia Figueroa, Nikolai Matni, Faizan M. Tariq. "Graph-based Path Planning with Dynamic Obstacle Avoidance for Autonomous Parking". presented at IEEE Intelligent Vehicles (IV) 2025. Project webpage

Farhad Nawaz, Faizan M Tariq, Sangjae Bae, David Isele, Avinash Singh, Nadia Figueroa, Nikolai Matni, Jovin D'sa. "Occupancy-aware Trajectory Planning for Autonomous Valet Parking in Uncertain Dynamic Environments". arXiv preprint arXiv:2509.09206

Darshan Gadginmath, Farhad Nawaz, Minjun Sung, Faizan M Tariq, Sangjae Bae, David Isele, Fabio Pasqualetti, Jovin D’ sa. "Active Probing with Multimodal Predictions for Motion Planning". presented at IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025. Project webpage