Gaussian Variational Inference Motion Planning
Leveraging control-inference duality for robust motion planning under uncertainty via Gaussian Variational Inference.
Projects
Covariance Control for Hybrid Systems
Optimal covariance steering for linear stochastic systems with hybrid transitions, applicable to contact-rich robotics.
Path Integral Control for Hybrid Systems
This project studies path integral control for nonlinear stochastic systems with hybrid transitions.
Path Integral Control for Partially Observed Systems
A sampling-based control framework for nonlinear stochastic systems under partial observability, leveraging path integral control.
Convex Optimal Control using Koopman Operator
Leveraging the Koopman operator and convex relaxation to solve infinite-time horizon optimal control problems with safety constraints.
Environment Representation in Sim-to-Real Imitation Learning
Evaluating the impact of sensor fusion and semantic abstraction on sim-to-real transfer for robot navigation.
Centralized Control of Micro-Drone Swarms
Implementing a centralized control framework for micro-drone swarms using external Vicon motion capture for state estimation.
Decentralized Field Coverage with Micro-Drone Swarms
Implementing a decentralized coverage control strategy on micro-drone hardware with collision avoidance and deadlock resolution.