People in the `star wars' cockpit. Defense system would require fast decisions by operators

If the United States ever deploys a ``star wars'' defense, its commander in some future crisis might have less than 60 seconds to reach one of the most fateful military decisions ever made. In the face of a suspected Soviet missile attack, a ``star wars'' system must be switched on and fired in that small amount of time, according to Pentagon figures. Any longer, and Soviet boosters would begin to rip through the all-important first layer of US defenses.

A human, not a computer, would choose whether to push the strategic defense button, pledge Defense Department officials. But they say it is not clear yet how much power people would have over strategic defenses.

``It is a tough problem,'' says Col. Richard Paul, Air Force Electronics Systems Division deputy commander for advanced technology.

Hardware that would be needed for ``star wars'' command has yet to be sketched. The computers, display screens, and complicated graphics that defense commanders would use are still vague concepts.

``To my knowledge virtually no work has been done on how this system is going to be controlled,'' says Dale Seastrom, manager of Strategic Defense Initiative programs for Ford Aerospace.

Missiles vulnerable when launched

The need for fast response by ``star wars'' defenses stems from the fact that shooting down attacking nuclear missiles in the first stages of flight is critical.

Just after launch, attacking missiles would be rising relatively slowly, their bright booster flames an easy target for the infrared sensors of defensive weapons. Each booster would be carrying multiple warheads, so the defense would get an extra payoff for each kill.

This vulnerable boost phase lasts about five minutes, at most. But US space-based interceptor rockets, the US weapons that would be firing back in any initial deployment of defenses, would take 2 minutes or more to reach their targets.

Thus ``star wars'' battle managers - both human and electronic - would have only a small slice of time to decide that an attack really was taking place, though satellite sensor data would probably be less than definitive. They would then have to tell numerous space-based weapons what their targets were, and order them to fire - all in seconds.

In fact, in 1982 when the private group High Frontier proposed a shield similar to today's plans for a first-phase ``star wars,'' it called for firing space-based interceptors before Soviets missiles were launched on the ground.

Today, an SDI time line calls for interceptors to be launched about 58 seconds after attacking missiles begin rising from their silos. But the document notes that US sensors will not register a Soviet attack until about 30 seconds after it has begun to take place, leaving 28 seconds for the US commander to decide to switch a space shield to ``on.''

That time window could be widened if the defense system used laser weapons, which would arrive at their targets at the speed of light.

Human control mandated

``Star wars'' presents ``a challenge to military commanders of a scope and speed never before encountered,'' concludes a SDI battle management study by a private contractor.

Critics charge that defense against nuclear missiles would be a sort of strategic ``beat-the-clock.'' They have long claimed that response would have to be so fast that the system would be automatically controlled by computer, without the overriding guidance of human reason.

As a result, a little-noticed amendment to the 1988 defense authorization bill, passed late last year, requires that humans would have to initiate the firing of any strategic defensive system. Informally, the provision was known to aides as the ``PTL'' amendment, for ``people that launch.''

SDI chief James Abrahamson, in a letter to Sen. Dale Bumpers (D) of Arkansas, said the amendment did not clash with Pentagon plans. ``Man will make the final engagement decision,'' wrote General Abrahamson.

Abrahamson also judged that ``boost phase performance degradation'' would be minimal if this engagement decision were reached ``between 40 and 100 seconds after launch'' of Soviet missiles.

Under most crisis scenarios, the commander of strategic defenses would have seen signs for days or weeks that an attack might be imminent - rising political tensions or movement of conventional military forces. Thus it is highly unlikely they would have only a few seconds to react to a complete surprise, Pentagon officials say.

In periods of high tension, the President or other national command authorities might delegate authority to start defenses down to the level of the system's military commander. In peacetime, politicians might have to give the go-ahead, although SDI researchers aren't sure how this would be done.

A flood of information

But beyond the decision to push the ``star wars'' button, it is not clear exactly what people would do in directing a defensive battle.

It would be impossible for a team of military officers to run numerous aspects of strategic defense, judges an SDI Organization battle-management report produced by the Armed Forces Communications and Electronics Association. There would be too much information flooding in, too much uncertainty, and action would be moving too fast.

An initial SDI system would be semi-automatic, with humans making some weapons allocation decisions, the report says. Eventually computers would make allocation choices as well.

The big challenge for researchers is duplicating in computer software the thinking process that humans would use in making these military decisions. ``We are actively working on the problem now'' with a human-factors working group of Air Force SDI researchers, says Colonel Paul of Electronics Systems Division.

Ford Aerospace, with hopes of future contracts, is spending its own money to build a prototype battle-management center in Colorado that can be used for SDI tests.

``What sort of information does an SDI decisionmaker need? Can a human turn on the system in time to work boost-phase defense? Do people have a place in weapons selection? How should this information be displayed? We're investigating those kinds of things,'' says Mr. Seastrom of Ford.

We want to hear, did we miss an angle we should have covered? Should we come back to this topic? Or just give us a rating for this story. We want to hear from you.