USARsim has been approved for use in a new Urban Search and Rescue competition for RoboCup. Because SIS has been the developer of USARsim this opens many opportunities for students here to get involved in either the competition or extending the simulation.

Pitt/CMU Virtual Robot Search & Rescue Team

Undergraduate-PhD- experience, fun, & publications

Team members- Fielding a RoboCup team requires people who are willing to develop interfaces, practice controlling robots, and doing all the things needed to keep the simulation running while finding ways to get to more victims, faster.

USARsim Development & Extension

Undergraduate-PhD- experience, fun, & publications

Telecommunications- Help develop and extend the simulation’s radio model

Psychology/Tele- Extend Jean Schultz’s work to associate radio degradation with video interference and behavioral consequences

Mechanical engineers: develop robot platform models using the Novodex physics engine. Novodex will be part of the UnrealEngine3 due to be released in late 2006. It allows multiple threads and will be hardware accelerated. We want to start development now so that we will be able to model tracked robots, etc. when it is released

Programmers: develop and enhance robot models & other parts of the simulation

Graphics programmers, artists, & GIS: help develop & extend models of the NIST campus and characterize and model damaged buildings for use in RoboCup

HCI: develop interfaces & conduct experiments with robot control.

Experiments with human control of multiple UAVs

Masters-PhD level research problem

Setup, develop, & maintain MatLab UAV models & OTBSAF simulation & interact with CMU research group & simulations. Design & run experiments to test interactive approaches to controlling UAVs

Requires:Experience with controls engineering, MatLab & simulation plus experience in experimental design and interest in human interaction with teams of autonomous robots. This is a great, ground floor area but you need to bring lots of specific skills & willingness to work.

General Framework for information fusion

PhD level research problem

This is a significant open problem that would make a great dissertation

I’m looking for someone interested in applying Situation Theory to interpreting sensor data "in context". The terrain analysis work is a great first step in providing a concrete way for us to get a context to "bootstrap" from.. Keith Devlin is still adjunct faculty & could probably be talked into helping..

Requires: Strong formal background, familiarity with logics, interests in philosophy, AI, semantics, etc.