Socially Aware Robot Navigation
Georgia Tech Research
Research on socially-aware navigation for Human-Robot Interaction using deep reinforcement learning on Turtlebot2 platform.
Read Paper (PDF)The Lab focuses on robust AI for sequential decision-making in environments under epistemic and aleatory uncertainty. We develop computational frameworks that integrate Machine Learning, Uncertainty Quantification, Reinforcement Learning, and Large-Scale Distributed Computing. Our research addresses the reliability gap in autonomous systems. We aim to architect agentic systems that maintain safety and performance guarantees in high-stakes, real-world deployment.
Optimization methods using gradient information for efficient parameter tuning in complex models.
Quantifying and reasoning about model uncertainty to make robust decisions under limited data.
Sequential decision-making algorithms that learn optimal policies through interaction with environments.
Forecasting methods for temporal data with applications in finance and system monitoring.
Probabilistic models for spatial data with applications in navigation and environmental sensing.
Autonomous navigation systems for robots operating in uncertain and dynamic environments.
Identifying unusual patterns and outliers in data for security and system monitoring.
Understanding cause-effect relationships for better decision-making and counterfactual reasoning.
Combining deep learning with RL for complex control and decision problems.
Understanding and classifying human activities from sensor and video data.
Georgia Tech Research
Research on socially-aware navigation for Human-Robot Interaction using deep reinforcement learning on Turtlebot2 platform.
Read Paper (PDF)Georgia Tech Research
An end-to-end method to train reinforcement learning agents using deep neural networks for partially observable Markov decision processes (POMDPs).
Read Paper (PDF)Georgia Tech Research
Addresses a major shortcoming to interval uncertainty approaches in computational mechanics by developing machine learning methods for interval field representation.
Read Paper (PDF)