Is This Missile Defense an Eagle--or an Albatross?

Next-generation satellites are the eyes and ears of Star Wars. But this glitchy program may not survive

Gunslingers in the Old West would jockey for a position with the sun at their backs. It's also a common military tactic: using the glare to mask an attack. Yet somehow, the problem of looking into the sun was neglected by the engineers who created a vital space-based sensor system for the Pentagon. The instruments could be blinded when peering at regions of the earth backdropped by the sun.

The system's contractor, Lockheed Martin Corp., knew of the problem five years ago--but underestimated its gravity. Then, with just a year to go before a scheduled 2002 launch, the aerospace giant overhauled the design and put "sunglasses" on the satellite sensors. "It's an exceedingly obvious thing that should have been compensated for," notes one dumfounded Defense Dept. insider. Correcting the oversight means a new round of tests, which will contribute to a delay in the launch until at least 2004.

For the program, called Space-Based Infra-Red System High (SBIRS High, or "sibbers," for short), the sunglasses episode was just one of many glitches. The program is now 70% over its revised $4 billion budget and 275% over the original figure of $1.8 billion. Because of the cost overrun, the Defense Dept. must cancel SBIRS High unless it certifies to Congress by May 5 that the system is essential for national security, that costs can be controlled, and that there's no alternative technology. Killing the program would be a huge setback for the missile-defense agenda: It would force the Pentagon to unveil a different, top-secret satellite scheme.

The SBIRS saga is hardly just another tale of a weapons program gone awry. This obscure program is a cornerstone of Defense Secretary Donald H. Rumsfeld's plans to improve missile defense and remote intelligence gathering. But now, E.C. "Pete" Aldridge Jr., former astronaut and current Pentagon acquisitions czar, is mulling whether SBIRS should get the ax. The Air Force and Lockheed Martin are feverishly arguing for its preservation, but neither would talk to BusinessWeek.

SBIRS wasn't controversial when it started. Everyone agrees that the current early-warning system, now three decades old, needs to be replaced. After several false starts, the Pentagon settled on SBIRS in the mid-1990s and selected a two-tier approach. SBIRS High would spot missile launches using six satellites, four in geosynchronous and two in highly elliptical orbits (diagram). And SBIRS Low, a constellation of 24 satellites, would handle the daunting job of distinguishing real warheads from decoys. Of the two, SBIRS High "was supposed to be the easy one," recalls Philip E. Coyle, a top Pentagon official in the Clinton Administration.

From the start, Lockheed Martin (LMT ) has had to contend with bureaucratic tweaking of the sort that has derailed many Pentagon development efforts in the past. For example, the Star Wars gang wanted SBIRS High to provide quicker launch warnings, and to analyze missile trajectory and other factors to determine whether a rocket was carrying a benign satellite or a warhead. Intelligence agencies insisted that the system should provide data about activities on the ground, including rocket-engine tests. And the Army fought for information about Scud and other short-range missile launches during combat.

Companies, meanwhile, battled over which technology would be best for SBIRS' mishmash of tasks. TRW Inc. (TRW ) argued for a straightforward upgrade of its existing imaging system, namely, arrays of detectors. These scan the earth in strips, which are assembled into infrared images. Rivals such as Lockheed campaigned for a different scheme, a "focal-plane array" much like what's inside a digital camera. It would look constantly at the whole earth through an open shutter, creating complete images. The end result was a complex and expensive compromise. SBIRS High would use a scanning array to detect areas of interest, then switch to a staring array to zero in on those specific targets. Unfortunately, says Theodore A. Postol, a Massachusetts Institute of Technology physicist, "nobody has figured out how to build a big enough staring array to do the job."

Another problem is the inherent conflict among various missions. For example, peacetime intelligence gathering calls for an extremely sensitive sensor to capture images of small, low-contrast details. But that kind of acuity in a missile-launch detector will produce a deluge of false alarms--especially since staring sensors have a high cry-wolf rate to begin with, says John E. Pike, head of, a defense-oriented advocacy group.

In addition, there's another type of false positive that remains a major thorn. To calculate the trajectory of a missile launch, two SBIRS High satellites send beams from their sensors toward the target and plot its position by triangulation. This works fine when there's only one target. But multiple targets produce more than twice as many ghost images because the beams intersect in other places. So if a rogue nation actually launched a multimissile attack, the chore of sorting out the real targets from the ghost ones could overwhelm SBIRS High's computers. That has already happened in one test.

Aerospace veteran Roy Danchick, a mathematician and former TRW senior staff engineer, claims he has devised a ghost-busting solution. It averages the independent tracking data of two birds, avoiding the problems of triangulation. In his tests, he says, the technique exceeded classified goals. But so far, SBIRS managers have shown no interest.

For the program's many woes, there's plenty of blame to go around. Uncle Sam repeatedly changed SBIRS High's requirements, and the Air Force and Congress kept tinkering with the budget, upsetting Lockheed Martin's plans. Analysts also blame a fix in acquisitions policy that was supposed to alleviate backseat driving by giving contractors near-total engineering responsibility. Several sources charge that Lockheed Martin, knowing there would be skimpy oversight, delegated the work to second-tier engineers. This issue also surfaces in a February report for the Air Force by an independent review panel. The Air Force vows rigorous supervision of SBIRS High--if it survives.

On top of all this, there are concerns about the partiality of one person who will help decide if SBRS High lives or dies. Air Force Under Secretary Peter B. Teets is a former Lockheed Martin exec who left after a 1999 management shakeup. If the project is killed, it could seem like sour grapes on Teets's part; yet if it's saved, that could appear to be due to Teets's former ties. To remove such clouds, some defense experts wish Teets would recuse himself. "You want these multibillion-dollar decisions to be made by the most objective official possible," says Joseph Cirincione, an arms-control expert at the Washington-based Carnegie Endowment for International Peace.

The betting is that SBIRS High will continue for now, but on a short leash. The program could swiftly die should it fail to meet future benchmarks. If SBIRS High does have to be scrapped, it would be a major blow to Rumsfeld's missile-defense goals. In that event, though, there is a fallback: a classified plan from the National Reconnaissance Office, the supersecret outfit that operates America's spy satellites. Rumsfeld's team would rather not play that card for fear of exposing too much of the NRO's shadowy world to the glare of publicity. So Rumsfeld has to be praying that SBIRS High won't get blindsided by still more troubles.

By Stan Crock in Washington

    Before it's here, it's on the Bloomberg Terminal.