The Georgia Robotics and Intelligent Systems (GRITS) Lab conducts research
in the general area of networked and hybrid control systems, with
applications to the control and coordination of mobile robots. The common
theme behind our different research threads is that we provide
theoretically sound solutions to practically motivated problems.
Research at the GRITS Lab is organized in three main areas:
A hybrid system is a dynamical system whose evolution
contains both a continuous and a discrete component. Examples of such
systems include switched systems, hybrid automata, and language-driven
systems. Our research in this area focuses on optimal control, observers
and observability, and specification languages for hybrid systems.
Although this research is mainly of a theoretical nature, the application
domains include behavior-based robotics, reactive and embedded systems,
and biologically inspired control systems.
Networked Control Systems
A networked system is one comprised of a
number of locally defined systems, held together by an
information-exchange topology. Examples of such systems include sensor
networks and multi-agent swarms. Our research studies how to produce
systems with provable global properties from local interaction and
communication rules. We are also investigating how such systems can be
programmed and controlled using tools from such different fields as
boundary value control and epidemic programming.
We have developed a number of experimental testbeds and
are conducting research on mobile robots and both structured indoor and
unstructured outdoor environments. Topics under consideration at the GRITS
Lab include how to make robots learn appropriate controllers as they
navigate through complex environments, low-complexity path-planning, field
robotics, and how to effectively structure and specify complex control and
mission-level tasks for single as well as teams of mobile ground robots.