In wrong hands, high-tech guns won't fire
January 7, 2005
By Anne Eisenberg
The computer circuits that control handheld music players, cell phones and organizers may soon be in a new location: inside electronically controlled guns.
Researchers at the New Jersey Institute of Technology in Newark are building a handgun designed to fire only when its circuitry and software recognize the grip of an authorized shooter.
Sensors in the handle measure the pressure the hand exerts as it squeezes the trigger. Then algorithms check the shooter's grip with stored, authorized patterns to give the go-ahead.
"We can build a brain inside the gun," said Timothy N. Chang, a professor of electrical engineering at the New Jersey Institute of Technology who devised the hardware for the grip-recognition system. "The technology is becoming so cheap that we can have not just a computer in every home, but a computer in every gun."
The main function of the system is to distinguish a legitimate shooter from, for example, a child who comes upon a handgun in a drawer. Electronics within the gun could one day include Global Positioning System receivers, accelerometers and other devices that could record the time and direction of gunfire and help reconstruct events in a crime investigation.
For a decade, researchers at many labs have been working on so-called smart, or personalized, handguns designed to prevent accidents. These use fingerprint scanners to recognize authorized shooters, or require the shooter to wear a small token on the hand that wirelessly transmits an unlocking code to the weapon.
At the New Jersey Institute of Technology, Michael L. Recce, an associate professor in the department of information systems, decided instead to concentrate on the shooter's characteristic grip. Recce created the software that does the pattern recognition for the gun.
Typically, it takes one-tenth of a second to pull a trigger, Recce said. While that is a short period, it is long enough for a computer to match the patterns and process the authorization.
To bring Recce's recognition software to life, Chang created several generations of circuits using off-the-shelf electronic components. He equipped the grips of real and fake handguns with sensors that could generate a charge proportional to the pressure put on them.
The pressure on the grip and trigger are read during the beginning of the trigger pull. The signals are sent to an analog-to-digital converter so that they can be handled by the digital signal processor. Patterns of different users can be stored, and the gun programmed to allow one or more shooters.
At first the group worked mainly with a simulated shooting range designed for police training. "You can't have guns in a university lab," Recce said.
The computer analysis of hand-pressure patterns showed that one person's grip could be distinguished from another's. "A person grasps a tennis racket or a pen or golf club in an individual, consistent way," he said. "That's what we're counting on."
During the past year, the team has moved from simulators to tests with live ammunition and real semiautomatic handguns fitted with pressure sensors in the grip. For five months, five officers from the institute's campus police force have been trying out the weaponry at a Bayonne firing range. "We've been going once a month since June," said Mark J. Cyr, a sergeant in the campus police. "I use a regular 9-millimeter Beretta weapon that fires like any other weapon; it doesn't feel any different."
For now a computer cord tethers the gun to a laptop that houses the circuitry and pattern-recognition software. In the next three months, though, Chang said, the circuits will move from the laptop into the magazine of the gun. "All the digital signal processing will be built right in," he said.
Michael Tocci, a captain in the Bayonne Police Department, recently saw a demonstration of the technology. One shooter was authorized, Captain Tocci said. When this person pulled the trigger, a green light flashed. "But when other officers picked up the gun to fire, the computer flashed red to register that they weren't authorized," he said.
The system had a 90 percent recognition rate, said Donald H. Sebastian, senior vice president for research and development at the institute. "That's better fidelity than we expected with 16 sensors in the grip," Sebastian said. "But we'll be adding more sensors, and that rate will improve."
Chang said the grip for the wireless system would have 32 pressure sensors. "Now, in the worst case, the system fails in one out of 10 cases," he said. "But we've already seen that with the new sensor array, the recognition is much higher."
Sebastian said the team was considering adding palm recognition as a backup.
To develop a future weapon, the university is working with a ballistics research and development company, Metal Storm, of Arlington, Va. "We'll use our recognition system on their weapons platform," Sebastian said.
The Metal Storm gun has plenty of room for the pattern-recognition circuitry. Rounds are kept in the gun's barrel, not in a magazine in the grip. There is a small amount of the gun's own electronic circuitry in the handle to control the firing, said Arthur Schatz, senior vice president for operations at the company. "Otherwise, it's pretty much empty, allowing the grip system to be housed within the handle," he said.
Captain Tocci of the Bayonne Police Department said the pattern-recognition technology was promising, particularly because accidental deaths occur when guns are not safely stored. "If a child picks up a gun that is not secured, this way it can't be fired," he said. Guns taken from a home during a robbery would be rendered useless, too.
"The premise the gun is based on has credibility," he said. When people see a live demonstration of the pattern-recognition system working, he said, "you think, yes, this is possible."
