Is missile defence a shot in the dark?

The US Congress is investigating claims of a cover-up over test results from the nation's missile defence programme. Owen Dyer looks at the controversial system

The advocates of "son of Star Wars" – a defence system to shoot down incoming missiles – claimed a couple of notable victories in 2001. Two trials ended with successful interceptions, bringing the tally to three direct hits out of five attempts. The explosions high over the Pacific Ocean sent hostile critics in Washington running for cover. The July test was timed to coincide with the appropriation of a further $8.3bn for the project, and the successful interception shot down any opposition that was brewing on Capitol Hill. "They hit a bullet with a bullet," said Trent Lott, the Senate Minority Leader. "We can develop that capability."

This year, however, has brought a sudden reversal of fortune. The US Congress's General Accounting Office is investigating claims that a scientific cover-up may have been perpetrated at the very heart of the missile defence programme. The programme's two prime contractors, the American aerospace firms TRW and Boeing, have been accused of manipulating data to hide the stark fact that their system cannot tell the difference between warheads and the decoys that accompany them.

The controversy dates from the first flight test in 1997, which the Pentagon said was a complete success. Although an interceptor missile was launched in that trial, it made no attempt to hit the dummy warhead. Rather, the mission was a fly-by designed to test the computer algorithm that recognises the target and the sensors on board the intercepting missile.

Scientists who have seen the flight data are said to be confused because vital seconds appear to have gone missing from each successive report. From a minute's worth of data released soon after the test, only 18 seconds were left in the 60-day review presented by Boeing, and these did not include the crucial last 11 seconds of the interceptor's flight. Moreover, of nine decoys that had accompanied the warhead, one was inexplicably absent in the figures.

The data fell into the hands of Professor Theodore Postol of the Massachusetts Institute of Technology, a long-time critic of "Star Wars" defensive shields. He claims to have found that during the missing seconds the interceptor was fixing on various dummies instead of the warhead. In fact, its attention was hopping almost randomly from one target to another. It appeared that the contractors simply chose those periods when it happened to be looking at the right target.

The dummies that accompanied the warhead were spherical "balloons" of various sizes. Unsure of the system's ability to discriminate the balloons from the warhead, the contractors had fed the dynamic characteristics of the objects into the system's computer. Unfortunately for them, one of the balloons failed to fully inflate. To the infrared eye of the interceptor missile, this object looked more like the warhead than the warhead itself.

Even after this awkward impostor was removed, they found that two other balloons were still preferentially selected by the discrimination algorithm. So they requested that the following tests include only four balloon decoys, a figure that was later changed to two.

In the end, last year's missile tests involved only one decoy, a spherical balloon seven times bigger than the warhead. Since the Missile Defence Agency tells the guidance computers that the warhead is the smaller of the two targets, it simply has to choose the dimmer of two bright infrared sources and make the intercept.

It's clever, but it tells us nothing about the system's ability to identify a warhead in a cloud of unknown objects. Philip Coyle, a consultant for the Center for Defence Information, and until last year the head of the Pentagon's Office of Operational Test and Evaluation, is worried about the system's reliance on information that just isn't available in the real world: "Obviously, our enemies aren't going to tell us which one is the real warhead."

The contractors promised a system that could reliably identify any warhead in any "threat cloud". Nira Schwartz, an engineer who worked for TRW, claimed before the first test that they could not come close to meeting this requirement. It was Schwartz's claims that led US Congressmen of both parties to demand an official investigation.

The Defence Department's Criminal Investigation Service (DCIS) also took Schwartz's claims seriously. They concluded that the discrimination technology "cannot and will not meet the contract requirements" and that TRW was "willing to repeatedly falsify documents submitted to the United States on this program". In a letter sent to the Missile Defence Agency, the DCIS investigators wrote: "There is no crime in producing a failed algorithm. The crime is in producing a failed algorithm and knowingly covering up its failure."

One might expect the Pentagon to take action, but instead the Army Department persuaded the Justice Department not to intervene, saying that its consultations with the DCIS led it to believe there were no grounds to proceed. In fact, the DCIS indignantly denies having been consulted.

Boeing, still the Lead System Integrator of the missile defence programme, refused to comment. A spokesman for TRW, which still holds the contract for the ground-based computer systems, said: "I can't comment on the congressional investigation, since the results aren't out yet. But the FBI looked into this and said there was no evidence of fraud."

The Boeing interceptor has since been replaced by a Raytheon model. Postol says this is no improvement. "The Boeing vehicle actually had better sensors." In any case, he says, no amount of technical tinkering could overcome the problem.

"Decoys are the Achilles' heel of missile defence," says Michael Levi, who studies the problem for the Federation of American Scientists. But the deadliest countermeasure is to put the warhead inside one of the balloons. "The empty balloons can be fitted with weights and electric heaters to mimic the loaded one," says Levi. "Better yet, make all the balloons different sizes and temperatures. There's no way to tell which one contains a warhead." This cheap and simple countermeasure, he says, cannot be defeated with today's technology.

The only response would be to shoot at all of the objects, but, says Coyle, "the system is really designed to launch one missile at a time". This raises another issue. Any chemical or biological missile would contain numerous bomblets to spread the lethal agent. Even without decoys, multiple bomblets threaten to swamp the missile defence shield. "The best decoy of all," says Coyle, "is another live warhead."

The missile defence establishment will be dreading the glare of publicity that surrounds congressional investigation. There is another flaw in their tests, one that came to light last summer. Every warhead intercepted so far carried a Global Positioning System (GPS) beacon that broadcasts its position to the ground computers. The Pentagon knows the beacon could be a public-relations disaster. "They want to get rid of it," says Coyle, "because it creates the impression of cheating, but it looks like we're stuck with it for the foreseeable future."

Now if they could just persuade the Iraqis and North Koreans to put beacons in their warheads, the "son of Star Wars" missile defence might actually work.