In the decade since the Federal Aviation Administration announced plans for a system to warn air traffic controllers of possible runway collisions, 59 people have died in five runway crashes while near-collisions have increased yearly.

But the FAA's major radar software program to help prevent conflicts between aircraft on the ground, originally scheduled for deployment in 1992, has been delayed at least another two years -- until 2002. And officials said it may never live up to some of its original promise, including warning of conflicts on taxiways or potential collisions with trucks and other support equipment.

"I think the FAA bit off more than it could chew," said Steven Zaidman, the FAA's associate administrator for research and acquisitions.

"I think it was too much."

The Airport Movement Area Safety System, known within the aviation industry by the acronym AMASS, has been burdened for years by a series of budget problems, lack of coordination and technical problems, according to internal documents.

Meanwhile, runway "incursions" -- FAA jargon for aircraft, vehicles or people bumbling into places for which they have no clearance -- continue to increase steadily, rising from 186 in 1993 to 325 in 1998. About 56 percent are "pilot deviations," while the rest are either "operational errors" laid to controllers, or vehicles and pedestrians who enter restricted areas.

This year alone there have been four major landing or takeoff incidents in which loaded airliners missed other aircraft by 50 to 200 feet.

The most frightening incident, involving two loaded Boeing 747s, occurred at 2:20 a.m. on April 1 at Chicago's O'Hare Airport. A China Airlines 747 crew apparently became confused and reentered the runway on which the plane had just landed. A Korean Air 747, taking off on that runway, flew directly over the Chinese plane at less than 50 feet.

The incident conjured up images of the world's worst aviation disaster, in 1977 at Tenerife in the Canary Islands, when two 747s collided, killing 538 people.

An FAA official, who did not want to be named, said AMASS would have warned the Korean plane in time to stop.

Federal Aviation Administrator Jane F. Garvey several months ago ordered that the FAA runway incursion program, including AMASS, be given top priority. But her lieutenants have concluded that necessary engineering changes will add two years to the latest plan to deploy and commission the system at 34 major U.S. airports.

Under the revised schedule, the first is scheduled to be commissioned at San Francisco in the summer of 2001, with all 34 airports -- including the Washington area's three -- commissioned by the summer of 2002.

In the past decade, AMASS has suffered from almost every known form of bureaucratic or technical problem.

At one point in the mid-1990s, for example, air traffic controllers and FAA managers seemed to be at war over whether the system could be used as a "snitch" to stain controllers' records.

Controllers also complained that engineers had designed the system with no thought about how controllers could use it and how its displays would be presented to them.

During most of 1993, 1994 and 1995, the air traffic section of the FAA stubbornly refused to validate the system.

Those days seem to be over. FAA officials, the National Air Traffic Controllers Association and various pilot and aviation industry groups are cooperating in developing a coherent runway incursion plan.

However, field tests in San Francisco, Detroit, St. Louis and Atlanta have shown that AMASS may never fulfill its planned capabilities. For now, AMASS will be set to warn only of potential runway collisions, such as one airliner cleared to land on a runway already occupied by another plane, or a plane entering a runway when it has no clearance. Taxiways and vehicles will not be protected.

The AMASS concept seems to be relatively simple. It is a software add-on to the Area Surface Detection Radar, known as the ASDE-3. This existing radar gives the controller an image of the airport with a small symbol marking each aircraft.

However, the radar has experienced problems with double images and blind spots, and AMASS has experienced multiple software problems and a nagging problem with false alarms.

"A little dinky system with only a processor and a radar is one of the greatest technical challenges we have," Zaidman said.

"This thing has to operate like a human brain."

"The two-year delay kind of hit me upside the head," said Bill Blackmer, head of the controller union's safety and technology department.

"Our representatives don't think it would take that long."

Although the controllers are supporting AMASS now, they are skeptical it will live up to expectations. The system is based on a radar aimed at the ground, meaning its signals can be blocked or scattered by almost anything, including buildings.

"Shooting radar at the ground is not the best idea," Blackmer said.

"Radar is very dumb. You send a signal there, it comes back here."

"What they've got is old technology," said Alexis M. Stefani, the Transportation Department's assistant inspector general for auditing, who supervised a critical inspector general report released in July.

Airlines, the FAA and the National Aeronautics and Space Administration are already experimenting with updated systems that could some day make AMASS obsolete, based on onboard technology that would report aircraft and vehicle positions, rather than a radar-based system that bounces signals off objects. Tests were conducted in 1995 in Atlanta and have just started in Dallas-Fort Worth.

This type of technology includes some aircraft systems that already perform other functions, such as collision avoidance. In addition, the Cargo Airline Association is testing a system that reports an aircraft's location based on global positioning satellite information.

Rep. Frank R. Wolf (R-Va.), chairman of the House Appropriations subcommittee on transportation, also is pushing the FAA to buy low-cost surface detection radar for smaller airports that do not rate the expensive systems going to 34 large airports.

Both Raytheon Co. and Dassault have such systems; the Dassault system was tested at Norfolk in 1998.

Wolf was angered by the FAA's reluctance to accept extra money for the system. He has asked Transportation Department Inspector General Kenneth M. Mead for an audit of the entire runway incursion program, possibly as a prelude to further congressional action.

Problems on the Runway

Fifty-nine people have died in these five runway incursions* since 1990 . . .

November 1996 Quincy, Ill. United Express/

private plane 14

November 1994 St. Louis TWA/Superior Aviation 2

February 1991 Los Angeles US Air/Skywest 34

December 1990 Detroit Northwest Airlines

(both planes) 8

January 1990 Atlanta Eastern Airlines/

Epps Air Service 1

*Incursions occur when aircraft, vehicles or people enter airport grounds without permission.

SOURCES: National Transportation Safety Board, Department of Transportation inspector general

CAPTION: Clearance Breach (This graphic was not available)