Embodied AI Systems
Integrating perception, planning, and control in long-horizon physical tasks.
Research Overview
Use this page to explain your themes, methods, and ongoing work beyond the short summary on the homepage.
Focus Areas
Research Interests
Robot Learning
Sample-efficient policy learning for manipulation, navigation, and adaptation.
Human-Robot Collaboration
Interfaces and autonomy strategies for safe, interpretable teamwork.
Edge Deployment
Moving modern autonomy pipelines onto real devices with tight compute budgets.
Experience
Research Experience
- 2025 – Present
Adaptive Mobile Manipulation for Cluttered Labs · Lead Student Researcher
Prof. Lina Zhao · Lab Seed Grant
- Building perception-driven manipulation policies that remain stable under shifting layouts and sparse demonstrations.
- Combined language-conditioned task prompts with vision-based affordance maps.
- Reduced reset time by introducing reusable evaluation scenes and scripted calibration checks.
- 2024 – 2025
Multi-Robot Inspection in Semi-Structured Environments · Research Engineer
Dr. Aaron Wu · Industry Collaboration
- Designed coordination logic for heterogeneous robots performing inspection rounds across changing indoor facilities.
- Implemented map-sharing and fallback routing for intermittent connectivity.
- Benchmarked task completion under degraded sensing and delayed communication.
- 2023 – 2024
Learning-Based Navigation Stack for Service Robots · Graduate Research Assistant
Prof. Lina Zhao · Graduate Fellowship
- Explored hybrid navigation pipelines that blend classical planners with learned scene priors.
- Added safety envelopes around low-confidence detections in crowded hallways.
- Improved recovery behavior after localization loss through state-aware replanning.
Interested in collaboration, advising, or internships? Point visitors to the Contact page.