ATLANTA -- Harold Greene stands in the middle of a shadowy alley, surrounded by red-brick buildings, sensing he's being watched. There's a guy with a gun out there, about to pop up from behind a dark window. Somebody moves. He draws. He shoots. Oops! He's killed a good guy.
But in this case, no harm's done. He's participating in a computer simulation of a crime scene in a laboratory at Morris Brown College.
In real life, such deadly tragedies occur all too often because of snap decisions made by police and others in high-pressure jobs requiring quick judgments.
That's why the military is funding a research center at Morris Brown, headed by Dr. Fernando Gonzales, a psychologist, to improve decision-making in life-or-death situations. Housed in a Victorian building on the college's downtown Atlanta campus, the Center of Excellence for Research on Training encompasses six laboratories that explore everything from pressure decision-making to training older drivers.
Gonzales says the multimillion-dollar Pentagon project, set up in 1993, not only will help prevent "friendly fire" accidents, but also will teach people in dangerous situations how to decide quickly when there's cause to shoot.
"Ultimately, this research will save lives and a lot of money," says Gonzales, 53. "Training is very expensive, and our methods are better and less costly." The military also wants to increase minority participation in psychological science, a field in which "minorities are egregiously" underrepresented, he says. The military needs more than a few good sharpshooters in an era when war-makers are often used as peacekeepers. It needs better ways to decide which jobs recruits are best suited for.
Better training is critical now more than ever, Gonzales says, because even "smart" technology can't prevent friendly fire casualties, which have killed Americans in every war. In the Persian Gulf, 24 percent of the Americans killed in action - 35 of 146 - fell victim to their own comrades.
Also, accidental shootings by police officers account for many deaths and injuries every year. One such tragedy in Atlanta occurred on Nov. 13, 1991, when Xavier Bennett Jr., 8, was killed by an Atlanta policeman involved in a firefight with one of the child's relatives.
Gonzales believes the techniques being pioneered at Morris Brown could save untold lives, but that they also have a host of other practical applications.
Many law enforcement agencies already use computer simulations of what psychologists call "shoot-don't-shoot" situations. But unlike designers o of commercial simulators, Morris Brown scientists are also studying specific psychological factors that lead to mistakes.
Volunteers are ushered into a dark rectangular room with black curtains on the sides and a huge screen on the far wall. They put on a mask that measures eye movements and "where you're gazing and when," says Greene, who also has a doctorate in psychology.
Subjects also are hooked up to electronic sensors that measure heart rate, respiration, blood pressure, body temperature and brain wave activity. The data also measure the effect of fatigue, concentration and excitability on judgment, says Dr. David Washburn, who is with the center and Georgia State University.
When all the data are analyzed, they point out strengths and weaknesses of each individual. For instance, heart beat rises as a person becomes excited, but that's good because excitement helps people focus. And brain activity measures fatigue, which can slow decision-making.
Washburn says the center's systems one day will allow individuals to work on their weaknesses. For example, a person who takes too long to spot a culprit with a gun might be taught to look at hands first instead of faces when sizing up a situation.
In the lab, subjects are given a small "weapon" with a button. They see images of red-brick buildings, streets and cartoon-like people who pop up in windows. If one is holding a rectangle, that means he's armed; a square indicates he's friendly.
The computer monitors how long it takes a subject to shoot, what factors went into the decision, and whether he killed a foe or an innocent person.
"If I shoot an innocent person, or fail to shoot at someone who shoots at me, is that a failure in perception? Did I fail to recognize it was a good guy or bad guy? Or is it a motor failure, or a judgment failure? We try to identify those component processes and see if those can be improved with specific kinds of practice," Washburn says.
The center's experts also plan to study ways to help police officers and soldiers overcome biases and prejudices. For example, in simulations involving black men and women, researchers found that black men had biases that women were less dangerous and didn't react as quickly to them as to men holding phony guns. The center also operates an "intelligent" Star Trek lab game capable of learning from its own mistakes. Scientists and students fight wars on computers designed to teach strategies and vigilance.
Tikado Shikano, 30, a Ph.D. candidate from Japan, says he's learned better ways to "kill Klingons" by anticipating common enemy strategies.
George Jones, 26, also a graduate student, says people learn better strategies the more they play, even when the Klingons' methods change, too.
"Human experience can make you better, but it can never make you perfect," he says.
Current simulators aren't capable of learning or must be reprogrammed to change scenarios, which is costly. The center's can be reprogrammed e easily, Greene says.
Washburn says the new tools have more applications than just for people in dangerous jobs. He sees the high-tech simulators as useful for people with attention deficit hyperactivity disorder, who have trouble concentrating.
"With our equipment, we can measure whether a person can concentrate for a short period or long period, how much distractions affect concentration, pinpoint those, and then give training on specific aspects they have trouble with," he says.
"There's a lot of promise here," Gonzales says. "It's going to be the new wave."