Developments to watch: Engineering can save train engineers

Control engineers should offer railroads simpler and more cost-effective systems to automatically slow or stop trains as needed to lower risk.


This antique “safety first” image was captured at the Illinois Railway Museum ( on Sept. 14, 2013. Some railroads still haven’t installed positive train control (PTC) systems to lower risk of collisions and derailments, recommended by the U.S.Are control engineers missing opportunities to more effectively lower risk for train engineers and their passengers? It seems that easier, less expensive retrofit applications of automation technology to trains could drastically lower risk of multiple accidents related to apparent errors by train engineers.

  • In Chicago on March 24, a commuter train crashed past a station bumper and went up an escalator, resulting in multiple injuries; had people been in the wrong place at the wrong time, deaths could have resulted.
  • In New York, a February 2013 crash killed four and injured more than 60.
  • In Spain, July 2013, 80 people died and more than 100 were injured when a train failed to slow for a curve.

Manufacturing machines can stop automatically to reduce risk. Automobiles can stop automatically to reduce risk. And trains? Some railroads have installed positive train control (PTC) systems. The U.S. National Transportation Safety Board has recommended PTC for years, said a 2007 post on the NTSB site. Current systems, some suggest, are costly and complex. Developments to watch: Smarter, less expensive, easier-to-install machine vision, sensors, and fail-safe controllers to slow or stop trains based on obstacles or conditions...before more people die. (This seems like so last century.) Putting related keywords into a browser returns attorney advertisements among search results. I cannot imagine that paying damages and attorney fees could be more cost effective than installing PTC safety systems on trains. Below, see related links.

- Mark T. Hoske, content manager, CFE Media, Control Engineering,

ONLINE extras

CBS Chicago: O'Hare commuter train crashes 

Huffington Post: NYC train crash 

BBC: Train crash in Spain 

NTSB said in 2007 it had been recommending positive train control for years

STEPHEN , NJ, United States, 04/01/14 08:28 PM:

In the past the PRR had a system that required the train engineer to acknowledge a signal / speed change indication with in 10 seconds the system used magnetic pickups and was in service for 40 years. So for all intensive purposes this is not a new idea and there should be no problem putting a refined version on line except for the cost!!
MICHAEL , CA, UNITED STATES, 04/30/14 10:19 PM:

Using industrial control components on a railroad has two technical advantages, as well as two cultural advantages.

The technical advantages are, first, the components are cheap. Compared to railroads, where everything is big, heavy, dirty and expensive, the factory floor is blessed with bargains. Given the low cost, it is possible to make redundant control sensors - say something to detect the presence, direction and speed of a train. And with a bunch of different technologies to pick from, the best choice is "all of the above". Seriously, using photocells, induction coils, pressure switches, laser gages, ultrasonics, etc., would produce a multi-fold set of data, that would initially feed indivdual PLC's. and then be processed by a larger computer control system.

Different technologies would produce different false positives, and different failure modes. With all systems working, speeds would be at the maximum, while each system failure would ratchet downthe speed. And some false positives - say a walking person interrupting a beam, would indeed send an emergency stop. It's hard to tell a tumbleweed from a trespasser, but erring on the conservative is always good.

The second advantage of shop floor technology is communications. The present showstopper for Positive Train Control is not the individual sensors,but the communications. The FCC has held up licensing the towers, to build the network. Shop floor communication - the two mile "microwave plug and play" - can be set up now, and as towers get permitted and built, the shop floor stuff can be used elsewhere. Plus, shop floor communications can use other nets - cell phones, satellites, etc. Just make it work. If it quits working, slow down til its fixed.

The cultural differences are worth looking into. Typically, a well automated factory usually runs at least two shifts, and possibly some or all of a graveyard/weekend shift. Control maintenance is done real time. If the line stops, the techs and engineers get right on it, and occasionally maintenance gets deferred to daytime (to contact a vendor, or work with a senior day shift engineer). But most maintenance means replacing a sensor module, or swapping a board or box. And the "macro" stuff sees troubleshooting on the bench, not on the factory floor. As such. the talent pool (which is hard to find) is concentrated on first and second shift.

On a railroad, maintenance gets done on graveyard shift. Boxes are often troubleshot in place. They are frightfully expensive, so spares are rare,or nonexistent. There is a high need for talent (even harder to find than in manufacturing), and a heavy concentration of a "graveyard culture". In some cases, thjis isolationist culture can lead to safety issues.

Having industrial components allows for "real time" replacement. Stop the train, have the tech jump off, swap the boards or box, jump back on the train. Maybe a five minute delay, instead of a slow four hour commute. A heavy inventory of spares. A solid daytime/swingshift concentration of talent, with graveyard coverage by late-stayers, or early starters - graveyard is split into two half-shifts.

My own experience of 40 years on the factory floor, and ten years on a railroad (driver, track controls, electric car technician) points me to this conclusion.
click me