I'm a PhD student at the GRASP Lab of the University of Pennsylvania, advised by George J. Pappas. Before that, I did my undergraduate studies at the University of Patras, where I worked with Stamatis Manesis and Anthony Tzes. During my PhD studies, I completed a M.S.E. in Robotics from University of Pennsylvania in 2020.
I've also worked as an Applied Scientist intern at Amazon Robotics and AI, where I applied state-of-the-art multi-modal learning methods for item identification.
My research interests lie at the intersection of robotics and artificial intelligence, specializing in the development of innovative autonomous systems and algorithms designed for optimal information acquisition and decision-making in dynamic, uncertain environments and in collaborative multi-agent scenarios. The focus of my current research is on building efficient task and motion planning methods leveraging Large Language Models.
Recent advances in metric, semantic, and topological mapping have equipped autonomous robots with semantic concept grounding capabilities to interpret natural language tasks. This work aims to leverage these new capabilities with an efficient task planning algorithm for hierarchical metric-semantic models. We consider a scene graph representation of the environment and utilize a large language model (LLM) to convert a natural language task into a linear temporal logic (LTL) automaton. Our main contribution is to enable optimal hierarchical LTL planning with LLM guidance over scene graphs.
We propose the Information-aware Graph Block Network (I-GBNet), an Active Information Acquisition adaptation of Graph Neural Networks, that aggregates information over the graph representation and provides sequential-decision making in a distributed manner. Numerical simulations on significantly larger graphs and dimensionality of the hidden state and more complex environments than those seen in training validate the properties of the proposed architecture and its efficacy in the application of localization and tracking of dynamic targets.
We propose a hybrid control architecture,
where a symbolic controller generates high-level manipulation
commands (e.g., grasp an object) based on environmental
feedback, an informative planner designs paths to actively
decrease the uncertainty of objects of interest, and a continuous
reactive controller tracks the sparse waypoints comprising the
informative paths while avoiding a priori unknown obstacles.
The overall architecture can handle environmental and sensing
uncertainty online, as the robot explores its workspace. Using
numerical simulations, we show that the proposed architecture
can handle tasks of increased complexity while responding to
unanticipated adverse configurations.
This paper addresses the problem of active information gathering for multi-robot systems. Specifically, we
consider scenarios where robots are tasked with reducing
uncertainty of dynamical hidden states evolving in complex
environments. We propose a novel distributed sampling-based planning algorithm that can significantly increase robot and target scalability while decreasing computational cost. We show
that the proposed algorithm is probabilistically complete and
asymptotically optimal.
Using a Voronoi-free area tessellation framework and a gradient scheme, we designed a control method for the coverage of a convex region by a team of mobile aerial agents (MAAs) under localization uncertainty. In this work, the agents are equipped with Pan-Tilt-Zoom (PTZ) cameras.
We designed a novel control scheme for the coverage of a convex region by a team of mobile aerial agents (MAAs) equipped with downwards facing camers, under positioning uncertainty.