Griffiss Air Force Base, N.Y.
If a ``star wars'' control room is ever built, it may resemble this secure Air Force lab. The lights are soft, the walls sound-absorbing, and the computers look like a new generation of video game. In a nuclear attack, such a center would have to watch thousands of objects: missile boosters, warheads in space, strips of foil chaff, decoy balloons. Still-unknown electronics would handle the task; today's technology, fast as it is, would simply crash.
Designing this battle-management equipment will be ``horrendous,'' says one scientist here. And it is such mundane-sounding problems that may determine the viability of ballistic missile defenses.
Lasers of gigawatt power and railguns are impressive technology. But if no control system tells them what to do, space-based weapons are nothing but man-made asteroids wearing American flags.
Strategic Defense Initiative (SDI) officials say the highest technical obstacles to missile defense include:
Computers. Computing hardware powerful enough to run a space defense now seems feasible, but scientists aren't sure if they can write the programs -- or software -- needed to make the hardware run.
Command and control. The various parts of a missile defense must be able to talk back and forth and work together, even in the face of massive attack.
Target spotting. A missile roaring out of a silo is as easy to see as a 10-story burning building, but cold warheads coasting through space are extremely difficult for sensors to ``discriminate.''
Power. A space-based weapon platform might require the energy of ``10 Hoover Dams in 1 second,'' says John Bosma, editor of the newsletter Military Space.
Transportation. The cost of putting things in orbit must be reduced 90 percent if space-based defense is to be affordable.
The men and women working on SDI say their job is to stand up sometime near the turn of the decade and say of these problems: ``Yes, we think they can be solved,'' or ``No, it's beyond us. Sorry.'' To make that decision, they must have some idea of what the entire missile defense might look like -- a task taking much of their attention right now.
In essence, SDI's system is being shaped by a brainstorming competition. Last year, 10 teams of companies won contracts to draw up an SDI ``architecture,'' or overall plan. This summer, companies such as as Martin Marietta, TRW, and Boeing were picked to polish their plans further, in the competition's second phase.
Those who've seen the closely held studies say three schemes for an initial missile defense system are emerging.
The first is multilayered, using weapons such as homing rockets on platforms in space and on the ground.
The second is less ambitious, featuring ground-based rockets and surveillance sensors that would be popped into space on notice of attack.
The third is a completely ground-based defense intended to protect Europe against intermediate-range nuclear missiles such as the Soviet SS-20.
None of these initial plans involve lasers or particle beams to shoot things down. Research in these exotic technologies has given the SDI program an otherworldly sheen and helped earn it the nickname star wars, but directed-energy weapons are still more prospect than fact.
Lasers and particle beams might be added to a defensive system after initial deployment, say SDI officials, particularly if the Soviets keep building new missiles. Eventually, the US might field a complex screen with directed-energy beams, kinetic-kill weapons, and as many as seven layers, according to SDI plans.
``It may be necessary to have directed-energy technology available in 2005, or 2010,'' says Navy Comdr. James Offut, with the SDI systems office.
The design of the SDI system is still evolving -- the company studies are more plans to make plans than plans in themselves. It's clear, however, that SDI is considering defenses more limited than a complex umbrella intended to be 90 percent effective.
``You can contribute to deterrence, to stability in the strategic sense, by constructing defenses less than thoroughly reliable,'' says Commander Offut.
For any US missile shield to be at all reliable it must have stalwart computers. And the computers must be fast -- so fast they would be to today's technology what an F-16 is to a biplane.
The Pentagon is counting on new semiconductor-chip design and new ways of linking computers together to provide this raw processing power. Even critics admit these approaches hold promise.
``The hardware program is not insurmountable,'' says John Kogut, a University of Illinois physicist who opposes SDI.
But teaching these speedy computers to operate is another matter. Mr. Kogut and other critics claim that software poses unsolvable problems for ballistic missile defense.
Everyone involved agrees that writing SDI's software would be a monumental task, the data-processing equivalent of building the Great Pyramids of Giza. A ballistic missile defense would need from 10 million to 100 million lines of software code that would tolerate faults -- ``understand what is a hiccup, and fix it,'' in the words of Offut.
Judgments of whether the job can be done at all depend crucially on new technology. To scientists working on SDI, a coming generation of software whiz kids will use new computer technologies, such as artificial intelligence (AI), which tries to duplicate an expert's thought process in software, to perfect strategic-defense computer programs.
``Artificial intelligence has had a lot of hype, but its applicability is real,'' says Ray Urtz, a technical director at the Air Force's Rome Air Development Center.
Air Force researchers using AI, for instance, are now developing a computer program that would help pilots pick the safest route through antiaircraft defenses to a target.
To critics, AI is akin to nuclear fusion power -- technology that marches bravely onward, but never seems to get anywhere. In addition, they claim that SDI software could never be fully tested without war, and therefore would not be trustworthy.
And some engineers say missile-defense programs would inevitably be full of conceptual errors; humans cannot foresee and write in computer code all the things that might happen in a nuclear attack.
``Sometime in my lifetime I might see something like this, but I'm skeptical,'' says Dr. David Redell of the Digital Equipment Corporation's Systems Research Center.
Missile-defense electronics running at full speed would have to keep an eye on all hostile missiles, warheads, and decoys; send orders to defensive weapons about what to shoot; and evaluate battle progress. To help this command-and-control process run smoothly, SDI scientists are trying to give the system's front-line ``soldiers'' as much responsibility as possible.
Surveillance satellites, for instance, might have powerful signal processors on board so they can process their own raw data. There might be ``lieutenant general'' computers in orbit, each capable of running the battle in its area of the front.
``That way, the enemy has to take out lots of things to wreck the system,'' says Dr. Charles Johnson, IBM director of battle-management architectures.
But SDI officials are not sure who, or what, might serve as a missile-defense commander in chief.
The problem is that a missile-defense system must have the reflexes of a soccer goalie. Its success might hinge on destroying intercontinental ballistic missiles during their a boost phase, which currently takes about 3 to 5 minutes. That does not allow much time to call the president in from a golf course.
So a US space shield would probably be controlled by an on-duty, high-ranking military officer, who would watch over a highly automated system, say SDI officials. Such delegation of authority is permissible, they say, because a missile defense system would likely not use nuclear weapons; set off by accident, defensive weapons would sparkle harmlessly in space.
``Although it might alarm the world, the consequences of a mistake are minimal,'' says Dino Lorenzini, head of SDI's pilot architecture program.
Advancing technology might at some point bring the president and other civilian leaders in on missile defense decisions, claims Mr. Lorenzini. ``They may have a little electronic gadget embedded in their ear at all times,'' he says.
Advances in many technologies will be needed if missile defense is to prove feasible. Lasers, computers, and communications are just a few of them. But one of the toughest technical problems, say a number of scientists, is an obscure one -- target spotting, or discrimination.
To see a rocket roaring up in boost phase all you need are binoculars and a relatively close seat. The trouble starts when the rocket burns out, and thousands of cold, dark warheads separate and go coasting across cold, dark space.
Critics, such as IBM scientist Richard Garwin, say that in this obscurity no electronic eye could reliably tell nuclear reentry vehicles from clever decoys. Military researchers are more optimistic.
``Finding them will require a multisensor approach,'' says Frank Rehn, a technical director at Rome Air Development Center.
Space-based radar might locate objects coming up through the clouds. Huge arrays of infrared detectors might pick up the trail, and lasers might push the objects, determining if they are warheads or lightweight balloons.
Much of the work on missile-defense sensors might be applied to other military missions. The Air Force, for instance, wants to use space-based radar to detect cruise missiles. Today's experimental ``stealth'' airplanes, intended to be almost invisible to current radar, might well look big as blimps to sensors developed by SDI work, according to scientists in and outside of government.
These fancy new surveillance eyes in the sky would require large amounts of electric power. In fact, generating power in orbit is itself a key problem facing SDI researchers. Around the SDI office, officials joke privately about running extension cords up into space.
Big battle-management satellites might need 75 kilowatts of power -- about as much as Skylab produced, says an SDI official. But the vast expanse of solar panels that Skylab used would be vulnerable in the heat of battle.
Exotic-weapons platforms would need even more electric power. Electromagnetic railguns might use short bursts of 1 gigawatt (1 million kilowatts); some types of lasers might need bursts in the 40-gigawatt range.
Among the power technologies SDI is looking at are chemically powered generators, advanced batteries, and small, orbiting nuclear reactors. The National Aeronautics and Space Administration, the Department of Energy, and SDI are working on the SP-100, a space reactor that might produce up to 2 megawatts.
``Even without SDI, this nation will have to face the political reality of accepting space nuclear power,'' says Air Force Col. George Hess, director of SDI's program in survivability, lethality, and key technologies.
It seems that any robust missile defense would inevitably count on some large components in space. They may be nuclear reactors and lasers; they may simply be large banks of infrared ``eyes.'' Somehow, these things would have to be sent into orbit, and that is yet another large problem.
It's a problem because currently transporting an object into space costs more than plating it with gold. The launch price for a satellite now hovers around $1,400 a pound. Unless that can be cut to $140 a pound or less, a space-based missile defense would probably be too expensive, SDI officials say.
As far as SDI is concerned, the space shuttle is only a pickup truck -- it can carry about 30 metric tons. The Fletcher panel, a Reagan-appointed group, headed by James C. Fletcher, which studied missile-defense technologies, concluded that SDI needs a fleet of big rockets able to boost 100 metric tons into low-Earth orbit and beyond.
In addition ``there's some advanced ideas for vehicles that take off from runways and then go supersonic [into orbit],'' says Gerald Yonas, SDI chief scientist. ``Those advanced ideas will come along in the fullness of time.''
The issue of cost, evidenced in the problem of space transport, hangs over the whole SDI program. Reagan administration officials have said that they would not favor deployment of a ballistic missile defense unless it is cheaper for the US to strengthen its shield than it is for the Soviets to increase their offensive forces.
Even if that condition is met, the overall price of the shield could still give Congress sticker shock. As a recent Congressional Office of Technology Assessment report notes, ``the cost and effort of a space-based [defense] does not end with deployment. Even in the absence of hostile action, there will have to be constant activity in space, occasionally with human presence, to maintain a working system.''
Would the system survive an attack? SDI officials say that might depend on a combination of things, such as physical shielding of space systems and tactics (satellites that dodge, perhaps).
``With a lot of Yankee ingenuity, I think we could build the lasers, build the particle beams to the required standards,'' says Cornelius (Cory) Coll III, leader of an SDI study group at the Lawrence Livermore National Laboratory. ``But this defensive system is more than a sum of its parts.''
If I walked into a garage and saw all these beautiful automobile parts, I wouldn't know when I put them all together if I was going to get a Mercedes or an Edsel. I think putting this together in a system is going to be the major challenge.'' -- 30 --