Team Performance in Human-agent Collaboration
One of the Army’s visions of a future combat system—soldiers, machinery and firepower—includes mixed teams of soldiers and “intelligent agents.” These agents could be robots or computerized support for decision-making.
The project’s purpose it to understand how humans and robots interact and, using this information, develop and evaluate technologies to improve battlefield performance.
IST formed a coalition of six Florida universities and a northwest Florida research institute to help the Army create a combat force of soldiers and robots. With over $3.64 million in support from the Human research and Engineering Directorate, U.S. Army Research Laboratory, IST will investigate how robots and soldiers might work together on the
battlefield. Partners with the university are the Institute for Human Machine Cognition, University of South Florida, Florida A&M, Florida State, University of Florida and Florida International University.
The results of this research could enable the early and effective deployment of high performance robotic teams, robots as members of human teams, and mixed human and robot teams for execution of complex tasks in dynamic environments typical of the battlefield.
While this project focuses entirely on military applications, the results will have broad-ranging implications in areas such as planetary exploration, health care, homeland security, air traffic management, and the control of complex adaptive systems such as information networks and energy distribution grids.
IST director Randall Shumaker is principal investigator on the year-long project. More information about the research and its importance to the Army’s visions of a future combat system is available from IST at 407-882-1300 or firstname.lastname@example.org.
IST partners extol virtues of living, learning and working in Florida.
It's no secret that to become a world-class center of excellence in simulation requires world-class components:
... A critical mass or related industry
... Academic and research commitments
... at the highest level
... An infrastructure that facilitates operations,
... communications and product movement
... A ready source of professional workers.
But did you know all those components are right here in Central Florida?
"They're ALL here in Central Florida" caught the essence of IST's 2006 partnership effort at the Interservice/Industry Training, Simulation and Education Conference (I/ITSEC) at Central Florida's Orange County
Convention Center Dec. 4 through 7.
Cosponsors of this year's effort were the Florida High Tech Corridor Council and Metro Orlando Economic Development Commission.
Daytona Beach Community College, Enterprise Florida and
Progress Energy were partners in the effort.
The exhibit's goal was to inform conferees, other exhibiting companies and simulation-related industry visitors to the show about the benefits of taking advantage of research opportunities, infrastructure, established high tech industry and workforce potential in Central Florida, the real simulation capital of the world.
Senior representatives of the cosponsors and partners were on hand during exhibit hours to answer questions.
See what some of our partners and sponsors have to say about the growing M&S industry in the Sunshine State...
Florida High Tech Corridor ...
Metro Orlando EDC ... Enterprise Florida
And read here how Industry Cluster, Academics, Infrastructure and Workforce have come together to create a world-class climate for the modeling, simulation & training industry.
Crowd behavior by the numbers
Here are two questions to ponder over your morning oatmeal: Can a mathematical equation predict crowd behavior? Should you care?
University of Central Florida mathematicians David Kaup and Thomas Clarke care, and so does the National Science Foundation, sponsor of the grant that funds their research. They and other IST researchers are striving to develop mathematical models and computer simulations of how crowds of individuals move around and interact.
And you should care. Results from Kaup’s and Clarke’s research could mean better designs for a building’s emergency exits, better crowd control devices at sports venues and a better understanding of how masses of individuals react to their surroundings.
To reduce a crowd to a mathematical model, researchers first must determine whether different crowds in similar situations reveal similar behavior patterns. These patterns of behavior, researchers have found, can vary depending on social forces within the crowd.
Snapshots of a non-panicking crowd before (far left) a rush to the exit (second left). The darkness indicates the “excitement level” of each individual. A few seconds afterward a group of pedestrians at the back breaks away, panicking and looking for a potential alternate exit.
Social forces are culture-dependent, say sociologists. For example, northern European people tend to be stand-offish and keep their distance, but Mediterranean people like to talk up close and have less regard for personal space. A major part of the research is adapting the models to incorporate such cultural factors.
Modelers must also consider a crowd’s excitement level, the result of each individual’s determination to get to some location, or away from something else. The collective level of intensity of each individual’s attempt to be anywhere other than their present location, modified by social inhibitions and the physical environment, determines to a great
extent the crowd’s “personality.” If researchers can find that crowds in similar situations display similar personalities, they can use that information to design crowd behavior models.
Simulation of the motions of individuals is one way to study crowd dynamics. IST investigators base their studies on a “social force model” introduced by Dresden University professors Dirk Helbing and Peter Molnar in the mid 1990s, modified by T. Vicsek in 2000 and further refined in 2005 by Kaup, former IST research associate (now assistant mathematics
professor at the University of Vermont) Taras Lakoba and UCF Simulation Technology Training Center deputy director Neil Finkelstein.
Kaup and the others refined the earlier model to more accurately reflect an individual’s movement within the crowd. Some of the earlier equations ignored many of the dynamics of “personal space” and allowed individuals to overlap or “crash” into one another with too much contact force.
To validate these mathematical models, the group is collecting and analyzing real world video footage of crowd behavior in non-emergency situations such as sporting events, concerts, schools, malls and church events.
FEST to test epidemic spread warning system
IST senior mathematician Dr. Thomas Clarke has designed a testbed to help validate a system designed to detect the spread of infectious disease.
The system, called BioDefend™ by its developers at the University of South Florida, has been tested in several Central Florida high risk facilities (theme parks, hospitals, a military base and a sports event) to determine whether “syndromic surveillance” can detect outbreaks sooner than routine surveillance.
Results of a two-year study (May 2002 –July 2004) reported to the National Centers for Disease Control indicated that the system can recognize outbreaks in near real-time.
BioDefend™ has been shown to work well with seasonal flu epidemics, but has not been sufficiently tested with epidemics of new diseases such as bird flu or with bioterrorism.
FEST, short for Florida Epidemiological Simulation Testbed, will mathematically simulate bioterrorism events and other outbreaks to test BioDefend™’s accuracy.
As part of his work in this area, Dr. Clarke this summer plans to teach a special topics course in the mathematics of epidemics.