Chasing the Jetsons

If life imitates art, why can't technology imitate cartoons?

By David H. Freedman|Monday, September 01, 1997
One thing you can say with some certainty about futuristic fiction is that it holds technology to be a double-edged sword. Soylent Green told us what happens when advanced health care keeps too many of us living too long: we end up eating our dead. Star Trek told us what happens when we travel out of our solar system: we get a whole new meaning for the term bad neighborhood. And 1984 gave us a taste of the fully wired society: imagine Bill Clinton listening in from your sock drawer.

All of which raises the interesting question, Is it possible to envision sweeping technological progress without attendant social, medical, or intergalactic ills?

Yes, it is, and for proof you need look no further than The Jetsons. That silly cartoon of the sixties posited a colorful future not noticeably more dysfunctional than the present, and yet it appeared to enjoy the benefits of technological marvels inaccessible to us. The world of the Jetsons had flying cars, robot housekeepers, food synthesizers-- well, actually, that’s about it, as far as major technological marvels go. But don’t knock it. Zipping airborne to the mall in a winged Hyundai, having your drapes vacuumed by a trash can on wheels, summoning up an instant three-course meal with a push of a button--what’s not to like about such a future? Especially if you can have these things without marauding aliens or post-nuclear global plagues. I mean, we’re talking awesome technology without a downside.

Okay, you’re sold. So when, you ask, is all this Jetsons stuff coming to Wal-Mart?

Winged Chevy

Flying cars have actually been around since the late 1940s. After all, anyone can take a small plane, pull the wings off, aim the thing down the interstate, and claim it’s a car. Some designs have been cleverer than this, of course, incorporating engines that can turn both propellers and wheels, removable wings, steering wheel-joystick combinations, and so forth. Not surprisingly, the price for this convertibility tends to be decent cars that are lousy planes, decent planes that are lousy cars, or more typically, lousy planes that are lousy cars too.

But convertibility isn’t the issue, argues Steven Crow, an aerospace engineer at the University of Arizona. The reason most of us pilot vehicles that leave the asphalt only when jumping potholes, he says, is that small planes--whether or not they convert into cars--are too expensive and too hard to fly. Plan on $150,000 for your bare-bones, cramped four-passenger plane. Add $3,500 and 60 hours’ worth of white- knuckled tutoring to get your pilot’s license. And don’t forget to factor in thousands of dollars for annual mechanical maintenance and other costs, and countless hours to plot flight paths. As for difficulty, think about how tricky driving through an unmarked parking lot can be, where you have just two dimensions. Throw in a third dimension, aerodynamics, rotations and gyroscopic phenomena, says Crow, and you’ve got a situation the human mind just doesn’t deal well with.

Crow thinks he can fix all this. His credentials are impeccable: He has a background in both aeronautics and aerodynamics, and--even better- -he doesn’t know how to fly a plane and can’t afford one either. If Crow can whip up something that works for him, then it should work for the rest of us as well. He’s been at it for eight years, and he considers it a noble calling. If the nineteenth century was the century of the train, and the twentieth century was the century of the automobile and large plane, then the twenty-first century may very well be the century of the personal plane, he says. After all, a small plane can cruise more than twice as fast as a car, get more miles to the gallon, and travel, well, as the crow flies. There’s nothing magical about the appeal of a flying car, he notes. I’m an engineer, not a cultist.

Crow and his students have over the years designed a series of flying cars--or roadable planes--called Starcars. Starcars 1, 2, and 3 were never built to actually fly, but Crow believes his Starcar 4, currently under construction, has a real shot at taking off, both literally and in terms of popularity. Starcar 4 is a three-wheeled vehicle with a front end like a motorcycle and a rear like a car. To keep its weight under 1,200 pounds, there are no doors; you lift the bubblelike canopy and step over the low sides to get in--a design not unlike that of the Jetsons’ Spacemobile, to which Starcar 4 bears an uncanny resemblance. A joystick controls Starcar 4 both in the air and on the ground. Each wing weighs only 50 pounds and quickly detaches so it can be hung on the sides of the vehicle when Starcar 4 is earthbound.

To keep costs down, Crow relies on a major technological leap: car parts. The trick is to make almost everything out of automobile components, he says. Only when you get to the wings do you depart from that strategy. In particular, high-performance car engines made of aluminum have gotten so light and powerful that they can take to the air. And they’re cheap. Whereas a real airplane engine would cost at least $35,000, Crow took a used car engine worth $495 and modified it--spending a total of $4,200 on the engine. As a result, his prototype can be built for less than $100,000, and Crow estimates a production version should come in at less than $40,000, maybe even as cheap as $30,000. I’d want this to appeal to the same people who buy bmws, Harleys, and Hobie Cats, says Crow. The Starcar 4 prototype will be drivable within a few months, he claims, and with any luck it will be flyable by 1999.

You might want to wait a few more years, however, until Starcar 5, now on the drawing board, hits your local Starcar dealership. Starcar 5 will have the added capability of carrying out perfect flights without the benefit of a licensed pilot. Equipped with two Global Positioning System radio receivers that independently track the vehicle’s position relative to satellites, as well as a gyroscopic system that can detect sudden changes in motion, this Starcar will know where it is at all times, how high it is, how fast it’s moving, and which way it’s pointing. Fed these data and given the destination airport, an onboard computer will be able to plot a flight path and then control the plane from takeoff through landing.

It’s actually not a radical concept, claims Crow. Pilots of large planes have already become somewhat secondary to the automated flight-control systems, he notes, and other researchers have already constructed and flown self-piloting model planes. Of course, complete self- piloting would be possible only if the national air traffic control system switched from the current centralized, ground-based approach to a distributed one in which planes broadcast their locations to other planes, and onboard computers figured out how to keep them away from each other. But a transition to such a free-flight approach is in the cards, Crow maintains.

Starcar 6, also in the works, will be even better. It will be a transformer vehicle with completely independent flying and driving components. When driven to an airport, the road module (two front wheels and power train) will quickly detach from the passenger compartment, and a sky module (propeller, engine, wings, and tail) can be popped on. When you land at your destination airport, you swap them back. Since consumers will need to purchase only the passenger compartment--they’ll rent the road and sky modules from either a private company or the flying-car version of a turnpike authority--the Starcar 6 should be as affordable as a Lexus.

Wouldn’t the Starcar’s very success, as manifested in a hypothetically huge number of users, cause the very same traffic jams in the sky that were the bane of George Jetson’s life? No way, says Crow. Despite all the nattering about crowded airports, the sky is virtually empty, he argues. And of the 50,000 airports in the United States, he says, fewer than 1,000 are currently in use.

Thus Crow seems to be promising an even better world than the Jetsons’. But there is only one shortcoming: in Crow’s vision of the future, you won’t be able to take to the air from your patio. Most of us don’t want others operating flying vehicles out of our neighborhoods, he explains.

People can just be so twentieth-century sometimes.

Team Rosie

Meet Rosie, a real-life robot named after the Jetsons’ wheeled maid and designed and built by a company called RedZone in Pittsburgh. She weighs 14,500 pounds, looks something like a very large sideways refrigerator on four wheels, and has a telescoping boom mounted on her top. She’s controlled by a human sitting some distance away--the distance being a good idea since Rosie earns her keep by excavating radioactive material from decommissioned nuclear reactors.

Rosie is typical of today’s state-of-the-art robot, in that a) she is designed to perform a highly specialized, somewhat technical task that humans are probably a lot better off leaving to something less than alive, b) she is dumb as a doornail, and c) she is pretty much the last thing you’d want serving drinks in your living room.

Why do robot labs tend to turn out such social misfits? Because no one knows how to build a robot that recognizes spoken English, doesn’t bump into furniture, can safely wield a dust rag around a vase, and for that matter, can even reliably distinguish a vase from a human head. Researchers used to think that because our brains seem to be able to handle such tasks effortlessly, it wouldn’t be hard to get a machine to perform them. Chess--now that seemed hard. It goes to show you.

Besides, NASA might be willing to pay millions for a robot capable of wandering around the surface of Mars and grabbing soil samples, but how much are you going to pay to avoid taking out the trash? As always, researchers tend to go where the money is.

But not all robot scientists have given up on the idea of a robot for the rest of us. Kazuhiko Kawamura, a professor of electrical engineering at Vanderbilt University in Nashville, is determined to see robots take their place in the home and office. When I started working on this ten years ago, no one else seemed interested, he says. I had to arrange some very creative funding. Now Kawamura runs the Center for Intelligent Systems at Vanderbilt, and the money is trickling in to support the development of a small line of what Kawamura calls service robots-- intelligent, autonomous robots that can perform useful jobs.

Kawamura’s current top-of-the-line robot is a dual-armed machine of vaguely humanoid styling, nicknamed, charmingly, Dual-Armed Humanoid. This robot has a sort of head, which consists of two swiveling video cameras mounted on a fixed platform. The robot can learn tasks by watching a human perform them and then mimicking the human’s actions. In this way, Dual-Armed Humanoid has learned, for example, to play the theremin, an electronic instrument from the 1920s that requires motions of the whole hand in the electric field between two electrodes rather than individual finger dexterity, of which the fingerless robot is in short supply.

To endow Dual-Armed Humanoid with the ability to grab small objects with its pincered arms--a surprisingly difficult task for most robots--Kawamura developed software that apes the strategy humans use to grab objects, at least when we’re infants. First a baby fixes the object in the center of his field of vision; then he sticks his hand out and notes where it is in his field of vision; then he moves his hand from the current location to the center of his field of vision, where the object awaits. Equipped with this fixation point strategy, Dual-Armed Humanoid can safely grab and handle silverware and can even spoon scalding soup from bowls.

Which is a good thing, because Dual-Armed Humanoid’s first intended application is feeding the sick, the handicapped, and the elderly in their homes. Since the robot is table-bound, Kawamura and his students developed a second, wheeled robot--HelpMate--with a single camera and arm that specializes in running through rooms to fetch items for Dual-Armed Humanoid. Thus HelpMate pops into the kitchen to grab a carton of orange juice and hands it to Dual-Armed Humanoid, which opens the carton, fills a glass, and holds it up to its master’s lips.

Why not build one robot that combines the capabilities of both? It’s too expensive says Kawamura. It’s much easier to build robots that specialize in certain tasks, and then get them to cooperate. We think in terms of robot networks, where the intelligence is distributed among two or more robots. As a bonus, Kawamura points out, a dual-robot system doesn’t necessarily become useless should one robot fail, as virtually all today’s robots do with impressive regularity. The second robot would be able to perform tasks on its own, such as picking up a phone and calling for help.

Kawamura’s group is also working on robots that can perform relatively complicated manufacturing assembly tasks, as well as a camera- equipped inchworm robot that can crawl along a ceiling on suction cups or magnets for police surveillance or for sneaking up on terrorists. But the big project on the drawing board is a more advanced version of the home- robot duo that will be capable of performing honest-to-goodness housekeeping chores, including cleaning up and taking out the trash. A sort of Team Rosie.

Kawamura says tests of Dual-Armed Humanoid with elderly and handicapped volunteers have been mostly successful, even though the robot tends to make a mess. One woman, however, complained that unlike her human helper, the robot couldn’t keep up its end of the conversation. I’ve thought about getting the group to teach the robots to crack jokes, says Kawamura with no hint of irony. But robotics graduate students aren’t very good at that sort of thing. Perhaps it’s just as well; after all, the Jetsons’ Rosie mostly used her dry wit to grouse. Besides, Dual-Armed Humanoid isn’t entirely without entertainment value, says Kawamura. It could play the theremin after dinner.

Mutable Feast

While we know that the Jetsons’ Food-a-Rac-a-Cycle instantly produced reasonably yummy-looking meals at the touch of a few buttons, the cartoon failed to specify the workings of this machine or the exact nature of its product. So we’re going to have to make some assumptions here. For one thing, the machine cannot have been materializing food out of thin air- -there has to be some limit even to what people of the future can do. It also could not have had access to thousands of ordinary ingredients, which it mixed and cooked in its microwave-oven-size chamber in less than a second. So we have no choice but to conclude that it must have been measuring out blobs of tasteless but nutritious paste, squirting in chemicals that approximated the flavor and color of the selected food, squeezing the paste into the appropriate shape and texture, and then flash- heating it to get the paste to set.

By coincidence, this is almost how food engineers produce many of today’s finest food analogues. If you didn’t know that food was engineered into stuff called analogues, that’s fair enough. But you can’t say you’ve never eaten any, unless you’re one of the few people fortunate enough to have managed to avoid artificially flavored fruit snacks, egg substitutes, veggie burgers, and hamburger extenders.

If anyone knows how to build a machine that could instantly whip up a passable analogue of virtually any food, it’s Susan Brewer. She is a food scientist at the University of Illinois at Urbana-Champaign, and for the past 17 years she and her colleagues have been searching for ever more efficient ways of turning bland paste into ever more convincing ersatz versions of an ever more varied selection of foods. There are ways to imitate almost everything, she boasts.

Dozens of foods--corn, whey, fish, wheat, peanuts, algae, and mushrooms--can easily be turned into bland goo, the basis of any analogue. But in the end, says Brewer, you’re probably going to go with the soybean. It’s cheap to grow, high in protein, and less likely than most of the other contenders to trigger life-threatening allergic reactions. What little flavor it has resides in the bean’s oil, so removing the oil and mixing the remainder with water yields a puddinglike slurry that’s about 95 percent protein--a gustatory tabula rasa. A lot of the bad associations people have with soy come from the late sixties, she says, when school lunch programs used a nasty version of soy that still had most of the oil in it.

Flavors are much trickier--some, such as coffee, result from the combination of as many as 600 different chemicals. In theory, of course, reproducing that flavor is simply a matter of throwing together the same chemicals. In practice it’s not that simple. For one thing, the components have to be in the right proportions, or the taste can be way off. Most of these chemicals taste bad separately, explains Brewer. It’s only when they’re combined in the right way that they add up to a nice flavor. To make things more complicated, we don’t really taste flavors other than sweet, salty, bitter, and sour; we mostly smell them, as molecules evaporate from the food in our mouths and travel up our back nasal passages to the top of our noses. So not only do you have to put together exactly the right chemicals in the right proportions, but you have to make sure they evaporate at the same time, and at the appropriate temperatures. That’s why serving temperature is so important to how foods taste. It’s also why food engineers often prefer to analyze the air around a chunk of food that’s been heated to serving temperature rather than a piece of the food itself.

Reproducing a flavor involves a lot of trial and error. The results tend to be better for fruit flavors, which are usually composed of lighter molecules that are easier to analyze and that evaporate evenly at lower temperatures. Meat flavor molecules, on the other hand, are heavier and difficult to identify in the lab, and they begin to evaporate readily only at temperatures above 120 degrees, which is just about as hot as the human mouth can tolerate. As an additional complication, almost all the flavor in meat resides in the fat, which people usually want removed for health reasons. And besides, fat molecules don’t mix well with the water contained in soy mixtures. So as a rough rule of thumb, a frozen artificial-strawberry-flavored soy-milk bar is likely to taste more authentic than a hot artificial-beef-flavored soy slab. But still, in theory, given the proper ingredients and proportions, a food synthesizer can squirt any flavor into the waiting soy paste.

All that’s left is giving the paste form. One method is to extrude it--to inject the paste between two long, threaded, revolving screws whose threads become more closely spaced as they spiral toward the far end. As the paste makes its way down the screws, it is mixed and squeezed more and more tightly, which causes it to heat up. By the time the paste emerges at the other end, it has already been cooked. Then it is forced--extruded--through nozzles. Adjusting the shape and size of the nozzles allows you to give the paste a bewildering variety of forms, ranging from the soft, stringy texture of shellfish to the firmer, grainy consistency of hamburger to the gooey, globular texture of pudding. You can do almost anything with an extruder, says Brewer.

As appetizing as flavoring and extrusion sound, many food analogues still need a bit of touching up before they’re ready for the table. Food engineers have worked some miracles with various chemical coatings, which provide that needed hint of stickiness or slipperiness. The artificial blueberry is one such triumph. The flavored, colored paste (usually made from algae rather than soy to get a more gelatinous consistency) is extruded into little globs and then dropped into a container of cold water mixed with calcium. The calcium molecules react with the algae to form a firm, fishnet molecular structure; once the structure starts to form on the outside of a glob, however, the fishnet blocks any additional calcium from getting inside. The result is a firm artificial skin around a moist jelly core.

Fresh vegetables, on the other hand, are problematic. They get their distinct crunchy-yet-moist texture from the microscopic cellulose- coated tubelike cells of which they’re composed. We’ve tried extruding tiny little spheres to imitate it, says Brewer, but it just ends up tasting as if you’re eating tiny little beady things. Artificial baked goods have also been a disappointment; no one has figured out how to get soy and most other bases to mimic wheat’s ability to cook into moist, puffy structures.

Over time, however, such holdouts are likely to succumb to relentless experimentation. Which means that the universal food synthesizer could possibly be more than just a dream. So how about it?

Yes, it’s possible, Brewer concedes, but I doubt anyone would bother to build one. Why would anyone pay goodness-knows-how-many thousands of dollars for a machine that would produce (at best) slightly weird versions of things they can buy for pocket change at the corner store? Even if you wanted to avoid meat for religious or health reasons, or to bypass foods to which you’re allergic, you don’t need anything as elaborate as a universal food synthesizer. And besides, real food is probably healthier anyway. Many people tend to think of analogues as being better for you because they may have less fat, says Brewer. But a lot of analogues do have fat, and some replace fat with sugar.

In other words, even if someone invented and marketed a universal food synthesizer tomorrow, it would probably end up in the back of thousands of closets next to the bread-making and cappuccino machines.

Still, it’s good to know the technology is feasible, should fresh food become unavailable. You know, if a postnuclear global plague contaminated the soil, for example, or if marauding aliens kept us captive in high orbit around one of the outer planets.

But then again, that’s a different sort of future.
Comment on this article

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar

Log in to your account

Email address:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »