Decision Making
Learning-informed Long-Horizon Navigation under Uncertainty. We present a novel approach to learning-augmented, long-horizon navigation under uncertainty in large-scale environments in which considering the robot dynamics is essential for informing good behavior. Our approach tightly integrates sampling-based motion planning, which computes dynamically feasible routes to the goal through different unexplored boundaries, and a high-level planner that leverages predictions about unseen space to select a route that best makes progress toward the unseen goal. Owing to its ability to understand the impacts of the robot’s dynamics on how it should attempt to reach the goal, our approach achieves both higher reliability and improved navigation performance compared to competitive learning-informed and non-learned baselines in simulated office-building-like environments.. (view the paper — IROS 2024)
Information Gain-based Pruning Strategy for Scout-Assisted Planning in Uncertain Graphs. A GNN policy over an information-gain formulation guides scouting drones, cutting a ground-robot team's travel cost by 30–45% across three environment types, with planning runtimes of a few seconds. (view the paper)
Learning-Guided Runtime-Prediction Motion Planning. Machine-learned runtime predictions guide a classical sampling-based planner's search toward low-cost regions of the state space — cutting travel distance 30% and improving planning runtime up to 3x for robots with dynamics. (view the paper — IROS 2022)
Nav2 Behavior-Tree Customization. Behavior-tree editing and TF/transform configuration to make fleet navigation decisions reliable in warehouse settings. (Bayro LLC — Dispatch Engineer, Mountain View, CA, Apr.–Aug. 2026.)