`Star wars' edges closer to deployment. Defense against ballistic missiles is entering a crucial phase, with US programs underway in six key areas. The Soviets are working on their own strategic defense system. But Moscow is demanding testing constraints as a precondition for arms control. First of three articles.

In its early years the Strategic Defense Initiative Organization lived like an exile. Its first home was an aging building far from the Defense Department's halls of power. Later, it moved to a cramped Pentagon space with movable partitions, old desks, and a temporary air. But today the SDI office is housed like a favorite son. Its new Pentagon headquarters are spacious and well appointed. Its lobby is sleek, gray, and busy with a stream of visitors - contractors, military officers, foreign embassy personnel.

Four and one-half years after it was launched by presidential edict, the United States project to build strategic defenses has gathered much bureaucratic momentum, of which its offices are only the most visible sign. In recent months, the Strategic Defense Initiative Organization (SDIO) has, among other things:

Won Pentagon approval to push six SDI technologies into the regular US weapons development process.

Established a ``Phase 1'' office, headed by a Navy captain, to plan for an initial deployment of defenses, perhaps by the mid-1990s.

To critics, all this represents movement but not technical progress. It is, they say, an attempt to institutionalize SDI so the next president can't tamper with it.

SDIO officials claim the moves are just logical steps in the long march toward a national decision on strategic defenses. ``The idea that this is some fast-track way to deploy a system is absurd,'' says Alan Mense, acting SDI chief scientist. Whatever the motive, it is clear that the SDI office is narrowing the focus of its work. SDI officials have long said that their job is to enable political leaders to make a decision about strategic defenses sometime in the early 1990s. With that date fast approaching, SDI money and manpower are flowing toward technologies that are furthest along in development.

The best examples of this process are the six technologies that made it into the Pentagon's weapons development pipeline by winning approval from a panel of Pentagon officials earlier this fall. They are the weapons and sensors that would form the basis for any initial defense, and by one estimate they will receive almost half the SDIO budget in coming years. Specifically, they are:

Killer satellites. These satellites would orbit the earth carrying clusters of interceptor rockets. Their job would be to shoot down both nuclear missiles rising through the atmosphere and warheads coasting through cold space in the mid-phase of flight. Much work still needs to be done on this weapon. The weight of rocket components must still be reduced significantly. Scientists must still see if the satellite ``can protect itself against hostile threats,'' according to SDIO documents released in August.

Ground-based rockets. These fast missiles, being developed by Lockheed Corporation, would be fired at warheads that were about to reenter the atmosphere. Ground-based missiles are perhaps the most mature SDI weapons - in a test of related technology this spring, the Army successfully shot a Lance tactical missile warhead out of the sky.

Surveillance systems. Three of the technologies now in development are sensors that would track the heat emitted by targets. These data would be used to aim defense weapons. Sensors watching missiles from launch through midflight would be based on satellites. Airplane and rocket mounted sensors would track warheads reentering the atmosphere.

Battle computers. A widely distributed network of computers, now being developed by the Army Strategic Defense Command and the Air Force Electronic Systems Division, would be used to command any strategic defense battle. This system would have to handle fast-flowing data from all SDI weapons and sensors. Producing the software necessary to handle this job is widely thought to be one of the most daunting technical challenges facing strategic defense researchers.

Earlier this year, high-level Pentagon officials voted to allow all these technologies to enter into what is known as the ``demonstration/validation'' part of the procurement pipeline. The panel concluded that SDI as a whole had made substantial progress, but that much work still needed to be done on the first-phase system. ``The design concept for a first phase is in an early stage and still quite sketchy,'' the panel report said. ``It takes the form more of a list of components than of a consistent design.''

Other reviews are still to come. But SDIO officials say passage over this first bureaucratic hurdle is a major event in the life of the program.

SDIO officials also say they are now focusing on three other key projects: construction of an SDI test center at an Air Force base in Colorado; choosing an ``integration contractor,'' who will oversee design of an initial defense system; and preliminary work on a large booster rocket that would be needed to lift SDI components into orbit.

To critics, the narrowing of SDI work is not the natural result of technical progress. They see it as a rush to judgment, a move to make the program politically unstoppable by getting something, anything, deployed as soon as possible.

Critics point to what the first-phase defense design does not include. There are only two layers of interceptor rockets, instead of the three planned. There is no directed-energy-beam sensor that would literally push at targets in space, to see if they were heavy warheads or light decoy balloons - another technology that was once considered for initial deployment.

And, of course, a first-phase defense would have no killer lasers. SDI scientists have long said such more exotic weapons would not be ready until after the turn of the century.

``The whole first-phase system might knock out only 20 percent of warheads launched in a Soviet attack,'' says Peter Zimmerman of the Carnegie Endowment for International Peace. Hundreds of nuclear weapons would be able to get through to the US, he claims.

SDI officials claim a first-phase shield would have a higher success rate than 20 percent. They also say that initial defenses could greatly bolster nuclear deterrence, as anything that makes Soviet war planners less confident about how many of their missiles might penetrate makes them correspondingly less likely to launch a first-strike attack.

``It doesn't take much of a defense to destroy that confidence,'' says Dr. Mense, SDIO's acting chief scientist.

One of the toughest problems that SDI faces is protecting the defense system itself. Critics say killer satellites would in turn be easy prey for Soviet antisatellite rockets or space mines. Mense acknowledges that survivability of a defense is a major issue. But he says that if the Soviets attacked a US space shield, it would be a sure warning that a nuclear missile attack was coming. Even if the defense were destroyed without firing a shot, the US would be able to disperse nuclear forces that would otherwise be vulnerable to surprise attack. ``If nothing else, you win time,'' he says.

But buying this time could be very expensive. Estimates about how much a Phase 1 defense might cost vary wildly - from $40 billion to more than $1 trillion. In any case, the price is likely to be high enough that Congress will have to be convinced an initial shield is valuable as more than a tripwire.

Legislators are already scaling back SDI's budget. This year, the administration requested $5.2 billion for SDI fiscal 1988 funds. Though the final figure has not yet been decided, Congress appears likely to reduce the SDI budget request by at least $700 million, and possibly by as much as $1.7 billion.

SDI officials say such a cut will push back the day a deployment decision could be made by two or three years, toward the late 1990s. Critics in Congress say the SDI schedule was too optimistic to begin with. ``They're kidding themselves if they think it will be ready to deploy in the early '90s,'' says a congressional staff member who closely follows the program.

There is a controversial shortcut SDI could take that might bring the day of decision closer. If the US read the Antiballistic Missile Treaty of 1972 more loosely than it currently does, SDI research could be restructured in a way that would shave about three years and $3 billion off the program, according to an SDIO report. The ABM Treaty prohibits deployment and advanced testing of defenses against ballistic missiles. Under a so-called ``broad'' interpretation of the pact, many SDI tests would be permitted because of a loophole that allows work on exotic new defense systems. This is the stance the Reagan administration favors. But important members of Congress say other parts of the treaty indicate that advanced SDI tests are not legal. So far, the White House has agreed to abide by this traditional, more restrictive treaty interpretation.

If the broad interpretation were to become national policy, SDI would schedule large experiments that would involve sensors, weapons, and battle-management computers working together. One such experiment would be Thor, a test involving tracking and destruction of a submarine-launched missile that could be ready around the turn of the decade, according to SDIO documents.

Absent a change to the broad interpretation, SDI tests in the next few years will generally look at space-shield components piece by piece. Next year, for instance, SDI will test infrared sensors mounted on the back of a modified Boeing 767.

The Soviet Union, for its part, says that what the superpowers should do now is negotiate precise definitions of strategic defense tests to be allowed under the ABM pact. Soviet negotiators in Geneva have given US counterparts a list of the technological limits - laser brightness, etc. - that they feel tests should not surpass.

Soviet leader Mikhail Gorbachev says that without progress on this front there will be no agreement to make large cuts in strategic nuclear arsenals. President Reagan has continued to insist that he will accept no further limits of any kind on the SDI program.

One little-noticed aspect of SDI work is the implication it holds for regular armed forces. SDI scientists claim that a number of advances they have made have direct application in more conventional weapons.

For example, a new gyroscope developed for space-interceptor rockets will be useful in the missiles carried by fighter planes, according to SDI.

Air Force Gen. John Piotrowski, head of US Space Command, says the space-based sensors that SDI is developing are needed for early warning of nuclear attack, even if the US decides not to erect strategic defenses.

Tomorrow: Pieces of the Soviet strategic defense puzzle.