Current Research Projects

The AstroNet: A Human-Centric Network of Free-Flying Space Co-Robots (Supported by NASA)

alt text 

NASA has long been concerned with the potential for micrometeoroid and orbital debris impacts to compromise the structural integrity of manned spacecraft. Visual inspection of spacecraft exteriors by crew members has been the preferred method of finding and assessing hull damage. In this proposal we aim at ultimately making the inspection and maintenance processes less time consuming and overall easier for the astronaut, and we introduce the AstroNet: A Network of Astronautical Free-Flying Co-Robots to interact with the crew members and assist them in their Extra-Vehicular Activities (EVAs) such as inspection, maintenance and repair of spacecraft exteriors. The Astronet is envisioned to: (i) safely surround the crew member during EVAs, (ii) perceive simple human commands (e.g., gestures) and interpret them into predefined tasks, (iii) respond to human commands by redistributing autonomously in space to dynamically and continuously improve task conditions (e.g., visualize areas beyond the line-of-sight of the astronaut, shed light, bring tools) in a human-centric way.

From High-Level Task Specifications to Geometric Control via Lyapunov Abstractions (Supported by Air Force Office of Science Research)

alt text 

The goal of this research project is to narrow the existing gap between high-level discrete task planning and low-level continuous control in complex multi-agent missions within a control-theoretic framework. We introduce the concept of a Lyapunov abstraction to enable the definition of a consistent mapping between high-level specifications and low-level control commands. A Lyapunov abstraction serves as a system model that by construction satisfies both the low-level dynamics and the highlevel goals, i.e., that captures the dynamics, tasks and interactions of a single agent with its environment, and is used as the unit element in the bottom-up composition of a hybrid system encoding the multi-agent mission. The Lyapunov abstractions we propose here are composed of high-level Lyapunov-like barrier functions (barriers) capturing high-level specifications and interactions among agents, and low-level geometric flows capturing feasible system trajectories. The main idea lies on the pairing of a Lyapunov-like barrier function and a geometric flow using notions and tools from geometric control and dynamical systems theory. The proposed method offers a reactive motion planning, decision-making and control design mechanism that is scalable with the number of agents and tasks, and thus applicable to large-scale systems involving hundreds of agents.