Is there a more absurd technology than positive train control, which Congress imposed as an unfunded mandate on railroads in 2008, and which supposedly would have prevented last week’s Philly Amtrak crash? Except it didn’t since its implementation has been draggy and its design so clearly inferior to cheaper, faster, more up-to-date solutions.
Congress enacted the law just after the first iPhone appeared but the conceptual underpinnings of PTC date much further back. And it shows. Railroads are already woefully behind a December 2015 deadline, and not just because of the cost and complexity (most of the technology had yet to be invented). An additional unmentioned factor is a lack of enthusiasm by railroads for what they know to be a burdensome, backward-looking, white elephantine approach.
Opinion Journal Video
At a cost of $13 billion or more, they must install tens of thousands of radio antennas along the track to inform a computerized locomotive about speed restrictions and other conditions ahead. A special challenge for Amtrak and other passenger lines has been obtaining the necessary radio spectrum. A three-year wrangle with the Federal Communications Commission and private license holders is why the system, already installed along the crash route, was not yet turned on by Amtrak.
Meanwhile, aboard every Amtrak train nowadays are hundreds of passenger smartphones constantly broadcasting their location using GPS over the same commercial airwaves that provide increasingly reliable call and broadband connectivity to the average consumer. Indeed, Apple and GoogleGOOG 0.32 % probably knew, or could have known, about the Philly accident before Amtrak did, based on the large number of phones that were moving and then suddenly not moving.
More than this, a train runs on rails and doesn’t have the opportunity to deviate or take wrong turns and detours the way a car does. So why not program speed limits of a given route into the train itself, as Google does with map information programmed into its self-driving car?
A Google car interprets and obeys traffic signals. Why not adapt the same technology to intervene if an engineer fails to notice or heed a signal telling him to slow down or stop? The complexity of railroad operations is significant, but not as complex as the decentralized operation of a hundred million cars on our roadways. So why not, in an emergency, allow a remote authority to take control of a train with a few keystrokes?
All this, of course, would merely be backup in case of operator error, but then the more convoluted and expensive positive train control also is offered as backup to a human operator. Nobody seems to be talking about getting rid of the engineer altogether, just protecting him from his own mistakes. Operator error is a fact of life: Still you have to ask why have an engineer, who doesn’t have to steer, navigate or worry much about other vehicles in his right of way, if he can’t be relied upon to apply the brakes for the curves along his regular route.
Positive train control is a classic example of what you get when Congress tries to dictate solutions to commercial and technological problems. Witness the Federal Aviation Administration’s perennial (and failing) struggle to catch up with the promise of information technology as applied to air traffic control.
Which brings us to another headline from the brave new world of self-driving vehicles. This month the truck maker Freightliner introduced a robotically controlled truck, licensed to operate on the roads of Nevada. Its onboard system, designed to relieve drivers of the monotony of motoring for hours down calm stretches of well-marked interstate, “never gets tired. It never gets distracted. It’s always at 100%,” company executive Wolfgang Bernhard told the media.
Alas, Mr. Bernhard deflated expectations by predicting that, though the system is ready to roll today, deployment is likely five years off. “The biggest obstacle that we see is the regulatory framework.”
Five years may be optimistic: An unspoken burden for the future is the legacy of the Toyota travesty of 2010, in which congressmen and, most damningly, a head of the Transportation Department, whose agency knows better, preferred to allege an undetected electronic bug in Toyotas rather than acknowledge that drivers (i.e., voters) cause accidents by pressing the gas instead of the brake.
This scandal, hugely costly to Toyota and largely fabricated, has never been acknowledged or investigated by the government or the media (which played a role). One big inconvenient precedent lies in its wake. As Toyota found, because it’s impossible to prove the nonexistence of a software bug, anytime there’s an accident involving a system in which software plays a role, the software will be blamed and the driver will be excused. Perhaps the only way forward, then, is to remove the driver altogether.