Fuzzy Logic Smart Machines

Mark C. Timney

Pop a snack into a fuzzy microwave over, leave the room and just forget about it.

No, your treat won't come out half raw or burnt to a crisp. Fuzzy microwaves cook foods perfectly all by themselves.

And no, they're not covered with fuzz. The word "fuzzy" refers to the fuzzy logic computer chip inside the oven. This chip turns an ordinarily dumb machine into one smart cookie.

Not Just Microwaves

Fuzzy logic is spiffy technology that can make almost any machine work better. Fuzzy washing machines automatically know how much soap and water to add to a load of laundry. They stop washing as soon as the clothes are clean. Fuzzy automobiles adjust themselves to fit driving conditions. This saves gas and gives a better ride.

Fuzzy cameras adjust themselves so even a beginner can take good pictures, with every photograph in focus. Fuzzy video cameras remove the jiggle. "It's a technology that will change forever the way we deal with machines," says Paul Freiberger, a technology expert in San Mateo, Calif.

Fuzzy Thinking

The fuzzy revolution started 30 years ago. Lotfi Zadeh, a computer science professor at the University of California, started comparing how humans and computers think. 9 "Computers depend upon exact information to do their work. We don't," Zadeh says. Most computers know only two commands: on or off. Humans know there are many steps in between. We don't, for example, turn a water faucet all the way on. We adjust it to put out just the right flow. That's a fuzzy judgment. Zadeh thought if he could teach a computer to handle fuzzy information, he would have a smarter machine. He was right.

The Fuzzy Difference

Let's say you want a computer to keep a subway train running at 30 miles per hour. A fuzzy computer would handle the job much differently than a non-fuzzy computer.

The non-fuzzy computer would need thousands of lines of programming like this:

If the train's speed falls to 27 m.p.h., then give the engine 10 percent more power.

It would need to be told how to react to every possible speed the train could reach. That takes lots of instructions.

A fuzzy computer's programming would be much simpler. Specific commands would be left out in favor of more general instructions:

If the train's speed falls a little below 30 m.p.h., then increase engine power a little. If the train's speed continues to decrease, then continue to increase power a little more.

As a result, the fuzzy computer would need only a few dozen lines of programming to do its job.

The Future Is Fuzzy

Japan's Sendai Subway is run by fuzzy computers. It is fast, fuel efficient and moves so smoothly that passengers don't have to brace themselves when it starts and stops.

Other fuzzy computers, or "expert systems," are helping doctors determine why a patient is sick, businessmen predict stock-market trends and engineers design bridges and buildings.

Computers are definitely getting smarter—and fuzzier.