An overview and update on the seven hit-to-kill tests conducted by the Defense Department between October 1999 and October 2002.

For decades, Kwajalein, a fly speck of an island in the middle of the Pacific, has been a U.S. missile test site. In tests of the midcourse missile defense system, this is where the military launches its "kill vehicle" -- a miniaturized package of sensors, computers, and thrusters -- which is designed to zoom up 140 miles, home in on an enemy warhead during the midcourse of its flight, and smash it to bits through the sheer force of the collision. The Pentagon calls this "hit to kill." It has been likened to hitting a bullet with a bullet, though at speeds of 15,000 m.p.h., the kill vehicle and the enemy warhead are traveling much faster than the speed of a bullet.

In the run-up to each test, the teams on Kwajalein ready Raytheon's $24 million kill vehicle, along with Boeing's $12 million booster rocket. At Vandenberg A.F.B. in southern California, 4,300 miles away, other crews prepare the $19 million Minuteman target rocket. Equipped with its mock warhead, the Minuteman rocket will simulate the flight of an enemy missile. At $100 million apiece, the hit-to-kill tests rank as the most expensive half-hour in military testing.

At least 20 developmental tests were originally planned, and each of these must succeed before operational testing can begin -- testing conducted by military service personnel instead of industry contractors and performed in realistic operational environments, at night, in bad weather, in conditions that would approximate battle.

Since 1997, the "kill vehicle" component of the Ground-Based Midcourse Defense (GMD) system has been tested eight times. In the first two tests, conducted in June 1997 and January 1998, the kill vehicle would simply "fly by" the dummy warhead without trying to hit it so that the kill vehicle's ability to locate the warhead and distinguish it from decoys could be evaluated. In the first, a Boeing kill vehicle was used; in the second, one manufactured by Raytheon. In both, the Pentagon declared the tests successful.

However, in March 2002, the General Accounting Office released a report that concluded that in the June 1997 test, the sensor used to discriminate between the warhead and decoys did not work as claimed -- allegations that had first been lodged by a whistle-blower (at TRW) and supported by the research of MIT physicist Theodore Postol.

The GAO concluded that the sensor systems on the kill vehicle malfunctioned and the sensor often detected targets where there were none. The report added that the contractors chose to analyze only about 12 seconds of data (out of more than 60 seconds taken) due to sensor malfunction and other problems. Toward the end of the test, the software began to incorrectly identify a decoy as the warhead. This data was eliminated in the contractor's analysis.

The first "hit to kill" or "flight-intercept test" -- that is, the program's first shot at actually hitting a mock warhead in space -- was conducted on Oct. 2, 1999. There have been six more hit-to-kill tests since then. Of the seven tests, five have been deemed successful by the Defense Department -- meaning that the kill vehicle was able to home in on and destroy the mock warhead -- but a fuller analysis reveals less encouraging details. In two of the successful tests, for instance, critical elements of the system failed. (See table.)

In addition, the earlier fly-by tests had included numerous decoys, but that had proved too difficult. So in the first five hit-to-kill tests, only one decoy accompanied the dummy warhead. It was a balloon much larger and brighter than the warhead, which made it easy for the kill vehicle to recognize. And it was easier still because a computer told the kill vehicle to look for the smaller of the two objects. In each test, in other words, the kill vehicle had been helped with the most critical technology of midcourse defense: being able to distinguish, in about 100 seconds, between a warhead and the swarm of decoys almost certain to be launched to disguise it.

"We are testing to learn; we are not testing as pass-fail for some operational reason," Lt. Gen. Ronald Kadish, head of the Pentagon's Missile Defense Agency, has said. In fact, all of the tests thus far have been "developmental" tests, designed to identify the strengths and weaknesses of the system's components, not "operational tests," which are conducted under realistic "combat" conditions. Philip Coyle, director of operational test and evaluation at the Pentagon from 1994 to 2001, and a consultant to FRONTLINE on "Missile Wars," says that "for all practical purposes, the first five flight intercept tests have been essentially the same." All seven tests, for example, were conducted with the mock warhead following the same flight course.

Some of the strongest criticism of the tests has focused on the controlled atmosphere in which the tests have taken place. Critics have pointed out that the kill vehicle is programmed with information about the size and shape of the mock warhead, information that would not be available under actual conditions. "That prior information," according to Coyle, "is the most important difference between early developmental tests and the later operational tests."

The table below describes the results of all seven tests since October 1999.

"HIT TO KILL" TESTS: October 1999 - October 2002

Updated: Oct. 15, 2002

home · the threat · the technology · the strategy · introduction · map
timeline · producer's chat · interviews · quiz: missile test · discussion
tapes & transcripts · press reaction · credits · privacy policy
FRONTLINE home · wgbh · pbsi

photograph © 1996 CORBIS; original image courtesy of NASA/CORBIS
web site copyright 1995-2005 wgbh educational foundation