The transit agency has reached “a milestone” in its effort to return to an automated system in which train operators do little more than make announcements, open and close doors, and keep an eye out for trouble (Rick Bowmer/The Washington Post)

Four and half years after a deadly rail crash forced Metro to let humans instead of computers drive its subway trains, the transit agency has reached “a milestone” in its effort to return to an automated system in which train operators do little more than make announcements, open and close doors, and keep an eye out for trouble.

Although the resumption of “automatic train operation” throughout the rail network remains a long way off, the work of preparing Metro’s busiest route, the Red Line, for computer-driven trains is “essentially completed,” said Rob Troup, the agency’s deputy general manager for operations. He called it “absolutely a milestone” in a project aimed at making the system more efficient and providing a smoother ride.

Getting the Red Line ready for automatic train operation is one thing; actually transitioning back to computer-driven trains is another. Troup said engineers have prepared a “very robust system-safety analysis” of the finished work. But the analysis still must be scrutinized by the National Transportation Safety Board.

“When we’re going to phase it in, we really haven’t made a determination,” said Troup, who is Metro’s No. 2 executive. “The thing we really don’t want to do is start putting dates out there until we’re certain the NTSB has approved it.”

For years, onboard computers controlled the movements of trains, using data transmitted by signaling equipment installed along Metro’s 106 miles of tracks. That made for smooth stopping and starting, among other efficiencies. Then, on June 22, 2009, a catastrophic failure of the track-to-computer communications process caused one Red Line train to slam into another near the Fort Totten station, killing nine people.

Since the crash, as Metro has worked to restore the reliability of automatic train operation, the agency’s more than 500 train operators have been at the controls, starting and stopping trains, accelerating and decelerating — obeying the same commands from the track-based signaling equipment that the onboard computers used to follow.

The difference: If a train operator, for example, were to receive a faulty command to proceed onto a stretch of track occupied by another train, he or she presumably would see the other train and disregard the command. An onboard computer, of course, does not have eyes. As it blindly followed just such a command in the 2009 Red Line disaster, a computer drove train No. 112 into the rear of train No. 214, which was stationary.

The operator of train No. 112, Jeanice McMillan, and eight of her passengers died in the accident, and about 80 other riders on the two trains were injured.

In fixing the system by which track-based signaling equipment transmits data to computers on trains, Troup said, Metro has upgraded or repaired many thousands of electrical components along the Red Line, “and we’ve proceeded to the Orange and Blue” lines.

As part of the process of returning to automatic train operation, engineers also have designed and built a unique electronic system, specific to Metro’s needs, that will alert train operators to any faulty communication between track-based signalling equipment and onboard computers. The NTSB required Metro to create such a system.

“Once the NTSB has made its opinion known on our system safety analysis, then we’ll really start looking at when we’ll go back to automatic train operation,” Troup said.

When it happens, riders are likely to notice. As train drivers, computers are far more precise and fluid than humans, whose skills at the controls vary widely.

A computer can keep a train moving at optimal speeds, smoothly accelerating and braking while precisely maintaining the prescribed distance from the train in front of it, then ease the train to a gentle stop at exactly the right spot along a platform. All of that makes for better on-time performance and more efficient use of the costly electricity that powers Metro’s trains, the transit agency said.

Humans? They don’t all drive the same way. Some speed up too quickly or slow down too abruptly, or both — giving strap-hangers the jostling sort of ride that daily Metro patrons have become familiar with since 2009, when the subway went to manual operation.

And there’s this announcement — commonplace nowadays — when an operator pulls up short at a Metro platform: “Stand back. Train moving forward.”

The role of automation in the operation of trains became a hot topic after a Manhattan-bound commuter train derailed Dec. 1 in the Bronx, killing four passengers. The man at the controls, a locomotive engineer for a decade, took the train into a curve at 82 mph, where the speed limit was 30 mph, authorities said.

The federal Rail Safety Improvement Act, passed in 2008, requires commuter and freight railroads and Amtrak to implement a computerized system known as “positive train control” by the end of 2015. If a train began traveling above the speed limit, the system would slow the train. In theory, positive train control would have prevented the Bronx derailment.

The rail safety law doesn’t apply to subway trains. Excessive speed normally isn’t an issue when a subway train is being operated by a computer. As for subway trains that operate in manual mode, such as Metro’s, electronic systems are in place that automatically enforce speed limits.

Some subways that were built long before the computer age — in Chicago and Philadelphia, for example — operate in manual mode. In New York’s subway, the nation’s biggest, computers drive trains on only one of the system’s 25-plus lines.

Among newer systems, the Los Angeles subway, opened in the early 1990s, also runs in manual mode. But the subways in Atlanta and San Francisco — which, like Metro, were built in the 1970s — use automatic train operation.

Although Metro was designed for automatic operation, the system has been running in manual mode for so long that a vast majority of Metro train operators haven’t experienced computer-driven trains. Of the agency’s 535 operators, a Metro spokeswoman said, 417 started their jobs after the 2009 crash, when automatic train operation was halted.

When computers eventually take charge of Metro trains again, those operators, after years of being in control, will be in a strange position. Imagine sitting, more or less passively, in the front cab of a crowded, eight-car rush-hour train — 400 tons of metal and humanity barreling along the rails at up to 65 mph — with little to do as an operator but surveil the tracks ahead while a computer console does the driving.

“I think, yes, the newer operators, who’ve only been in manual mode, are going to feel some apprehension at first,” said Jackie Jeter, president of the union representing Metro’s train operators. “Even if you’re an old operator like me” — that was Jeter’s job in the 1980s and ’90s — “that accident in ’09 shook your confidence in what automatic train operation is able to do. I mean, I couldn’t have fathomed that happening.”

For an operator, placing life and limb in the hands of an automated system that has failed tragically “will take some getting used to,” she said. “But we’ll get there.”