Blind man at the brake
Sep 19th 2008
A little automation could save lives
TIME and again, evidence shows computer-controlled vehicles have fewer accidents than those driven by people. Still, we have a pathological reluctance to surrender control. We seem to find all-too-frequent deaths caused by human carelessness somehow preferable to the obscure possibility of death by bloodless automaton.
The rail crash on September 12th—America’s deadliest in 15 years—would have been wholly preventable had an electronic controller been allowed to intercede. Instead, a Metrolink commuter train heading north from Los Angeles with 220 people on board collided head-on with a southbound Union Pacific freight train, killing 25 people and injuring 135 more.
The Metrolink’s driver, who died in the accident, ran through three yellow and red signals warning of an approaching train, and overrode a switch that would have diverted one of them. Investigators have subpoenaed his mobile-phone records, following reports claiming he was texting trainspotters just before the crash.
Metrolink has one of the worst safety records in America, chalking up 74 fatalities over the past decade. Only New Jersey Transit, with 80 fatalities, is worse—but at least New Jersey’s trains cover four times the miles and carry five times the passengers annually than trains in Los Angeles.
Providing automatic train controls that warn a driver who fails to notice a signal or exceeds a speed limit—and then take command if the warning goes unheeded—is not rocket science. Airliners have had collision-warning and prevention systems since the mid-1980s. Providing them for a fixed network on the ground should be even easier. The National Transportation Safety Board has endorsed a technology called Positive Train Control (PTC), but only 4,000 of America’s 140,000 miles of track use it.
Amtrak trains on the Northeast Corridor between Washington, DC, and Boston have been upgraded. Union Pacific has done likewise on a stretch of track between Chicago and St Louis. New Jersey Transit is installing a PTC system on some of its commuter lines, and Alaska Railroad is a testing a PTC system that will eventually be introduced throughout the state.
At its simplest, PTC relies on three basic components—equipment in the cab, a control centre elsewhere, and a wireless link between the two.
A sat-nav receiver in the cab gets the train’s precise location from signals broadcast by GPS satellites overhead that are refined by positional data from a network of ground-based transmitters. This so-called Nationwide Differential Global Positioning System can pinpoint objects on the ground to within one metre, compared to anything up to 20 metres for conventional GPS.
Using the wireless link in the cab, computers at the control centre collect location data from all the trains on the network and issue “limits of movement authority” (LoMA) and speed limits for each individual train.
Back in the cab, an on-board computer compares the PTC-dictated LoMA and speed limit to the train’s location and speed. The computer warns the driver when the train approaches its speed limit or nears the border of its LoMA. If the driver takes no action, the computer applies the brakes.
But PTC is more than just an emergency brake and messaging system. It can “pace” the engine to adjust the headway between consecutive trains, thereby increasing the capacity of the line. One of the largest benefits of PTC is a reduction in equipment cost, thanks to a 5% to 10% increase in the speed it allows freight cars to travel.
In addition, PTC can keep an eye on wayside equipment like track switches, and even take control of them if disaster looms. It can also tell the control centre how the train is performing, and diagnose the wear and tear on the locomotive’s components.
It’s hard to see why the majority of American railways resist PTC. Apart from providing a greater level of safety, PTC will save them money. Pilot tests done a decade ago show PTC saves at least 5% on labour costs alone—a $250m annual saving for the industry as a whole. Meanwhile, speeding up shipments is expected to bring the railways an extra $560m a year.
Sure, the technology is far from mature. Signals can get lost in tunnels or narrow canyons. But the technology has improved greatly since it was first mooted more than 30 years ago. And it’s got cheaper, too.
Still, the railways balk at the cost. They fret that each of their 20,000 front-line locomotives would need to be equipped with GPS, an on-board computer and a data link. They would also have to invest in radio repeaters and other trackside gear for monitoring switches. And additional hardware and software would have to be bought for the control centres.
The overall cost, says the Federal Railroad Administration, could range from $2.3 billion to $4.4 billion spread over five years, depending on whether the installation was an “overlay” on existing signalling systems or a completely new system. But once it was up and running, the industry could be reaping benefits to the tune of $2.2 billion to $3.8 billion annually. That’s an internal rate of return of 44% to 160%, say consultants.
Meanwhile, foot-dragging by the industry has stalled legislation that would require PTC to be adopted nationwide. A bill proposed by the House of Representatives requires America’s main railways to install PTC by 2014, but does little to ease the financial burden. By contrast, the Senate bill would not take effect until 2018, but focuses on heavily congested areas—like Los Angeles where freight and commuter trains share busy single-track lines and have to contend with lots of level crossings.
Railway safety officials have identified almost 1,000 accidents over the past dozen years that PTC would have prevented. The time has come, say critics, for the congressional timetable to be shortened, and for the government to come up with matching funds so PTC can be installed before another deadly accident occurs.