Zsolt Kira

Site Map | Home
Bio | Resume | Schedule | Public Key
Statement | Knowledge Sharing | Sensorimotor Integration | Multi-Robot Systems | Case-Based Reasoning
Software Released | Linux on Laptop
All Albums | IROS & EpiRob 2007 | Burning Man '06 | Europe '05 | Japan '04 | Nantahala '03
Bio | subglobal8 link | subglobal8 link | subglobal8 link | subglobal8 link | subglobal8 link | subglobal8 link

Research Statement

I believe that learning, adaptation, and development are important concepts in the creation of robots that act intelligently.

Learning and adaptation refers to the fact that in order to make robots that are useful in many situations and tasks (as opposed to a really niche robot), they must learn how to perform tasks and adapt to changing situations when they arise. Although there have been many methods involving hand-coding of contingencies, I believe robots should learn these through interaction with the world.

Development refers to the fact that robots must attain increasingly sophisticated capabilities through interaction with the world. This is a recently recognized desirable property, and conferences such as Epigenetic Robotics and ICDL have focused on it.

Main Projects

Heterogeneous Knowledge Sharing
This project is about sharing of knowledge among heterogeneous robots that may differ in their perceptual or motor capabilities, ontologies, or experiences.

Two Heterogeneous Robots

Sensorimotor Integration
This project seeks to investigate statistical models of correlations between sensing and action in order to detect changes in the capabilities of the robot.


Multi-Agent Robotics
Anytime Learning uses genetic algorithms to find increasingly more effective control system for a task during the actual performance of the task. When aspects of the task change, a task monitor detects this change and in turn updates the simulator used in learning. I adapted this framework to multi-agent systems where robots can share their knowledge of the state of the world.


Case-Based Reasoning
The work done in this lab applied Case-Based Reasoning in order to learn behavioral parameters used in a behavior-based system.  As part of the MARS project, I added a memory management (or case library maintenance) system to this existing framework whereby cases could be removed if a new situation arises that is not in the current library, but the library is full.


Other Project

Visual SLAM
We investigated use of Visual SLAM in an outdoor environment, in addition to sharing of the resulting topological map among heterogeneous robots. Project webpage is here.



Teleoperation
During Fall 2002, Jung Hoon (another graduate student) and I implemented a user interface for teleoperation. It displays hardware information (e.g. battery) as well as communications quality (ping time) and several other features. This work was part of the FCSC project.

Robotic teleoperation interface

©2007 Zsolt Kira