Statement of Grady Cothen, FRA

MR. COTHEN: Thank you, and good afternoon. It's good to be here representing the Federal Railroad Administration. The topic today draws me into a point of tension with a National Transportation Safety Board recommendation, but before being sucked into that vortex, I would like to -

MR. GOGLIA: Now, careful. You're close to some of the heavy weights around here. [Laughter]

MR. COTHEN: I'm already being cautioned. Before that happens, I would like to affirm the commitment of the Federal Railroad Administration to the use of high-quality data recording in order to advance the cause of railroad safety.

As Ed English has told you previously in this conference, we're working with the Board and others in the industry on a rulemaking for next generation event recorders on locomotives.

Increasingly in the railroad industry, as elsewhere in transportation, we have available to us a great amount of data from sources, such as signal inter-locking controllers and grade crossing warning device controllers. In the operation centers, from centralized traffic control systems, computer-aided dispatching systems.

For Class 1 railroads at least and for some of the regional railroads, communications between trains and dispatching centers are recorded from the railroad radio service road channels, and, of course, there are a variety of other sources of data which are available to the Board and to the Federal Railroad Administration for accident investigation purposes, and there's an excellent spirit of cooperation, and railroads are forthcoming in providing data, both the data that is required to be maintained and data that is maintained for other purposes. That's the good news.

The bad news is that I think in the railroad industry, we're not doing as effective a job as we could in terms of utilizing all the data that is on hand, and I've been suitably impressed with the other presentations here today that point to the direction that we need to go.

Without question, accident investigation is precisely the wrong time to be asking questions regarding performance of people and systems. We need to be and hopefully we will begin to be more alert to the use of data systems in working towards safety assurance, early hazard identification, quality control of systems, and as we look at the potential for varying inspection, testing and repair intervals for equipment and systems, use of data to drive reliability assessment processes and therefore appropriately-scaled attention to safety needs.

We need data sometimes to enforce our regulations as well, and we can't neglect that; nor should we believe that in every case, the data that is produced has to be available for that purpose. There are good reasons for it not to be.

Let me talk, though, about, in addition to public needs, private expectations because that's kind of the theme of this panel, and if I were on a different panel, I promise I would be talking about other issues, but there are legal and labor relations kinds of issues, public policy issues that swirl around the increasingly-capable electronic systems that we have available to us today.

When I use the term "expectations", I really don't intend that to be a loaded word. Some of our private expectations are reasonable, and some of them are otherwise. Even expectations that are reasonable today are often capable of being altered as a result of changes in public policy, and anyone who doubts that can go back to the beginning of the discussion when Jim Burnette started talking about alcohol testing in the railroad industry and see how far we have come since that day.

The fact, however, that we can make the journey does not in every case mean that we need to make the journey or should make the journey. So, we have to work through all these various issues.

Private expectations include corporate expectations. Railroad carriers, passenger and freight, public and private sector, are paying for these systems. They certainly have a reason to be free from arbitrary requirements. Just because somebody thinks it's nice or it would be interesting or because it would make a nice study at a university some day clearly is not a sufficient reason to issue a requirement.

Certainly our transportation companies need to have the ability and some degree of predictability to deploy systems that meet their business service needs, and they're suitably configured to do that without believing that in every case, the regulator's going to jump on them and say all right, now, we'll freeze that, and that has now become a regulatory requirement.

The effect of that kind of practice will be to chill the use of innovative approaches to not only capturing data but analyzing data. I think that probably backhandedly, we have done a little of that already in the railroad industry.

Certainly corporations have reason to complain that abusive discovery related to litigation can occur. That doesn't mean that's the case every time, and normally we trust the courts to police that. Nevertheless, it is the case that individual data points can sometimes be taken and represented in a way that corporations may view as unfavorable to their interests. So, we have to deal with that expectation-a negative expectation in this case.

Individuals as well have their expectations. Fundamental fairness is one of them. We all want to be evaluated--whether we're caught in a radar trap or whatever it may be--we all want to be evaluated by accurate data, data that's technically correct, and also data that is representative in terms of the parameters that are reflected there.

You know, it's possible, for instance, for a locomotive engineer to get into an over-speed violation, but the reason for that may be that there was a failure of dynamic brakes in the consist and time taken to recover from that, and if one is known but not the other, then there's a serious disadvantage to the individual there.

We all like to engage in conversations in the work place, I think, that probably are not, strictly speaking, business-related. That's going to be particularly true in the railroad industry where you have crew members working up to 12 hours a day over long distances with a significant amount of down time, standing in sidings waiting for passes, and the reality there is that there is a human price to pay if folks don't feel like they can speak their mind.

At the same time, obviously, we all need to be "on task" when our employment requires us to be "on task." So, the topic area takes me into an issue that we have under a pending National Transportation Safety Board recommendation that is specifically that we have a cab voice recorder. The recommendation emanated from the Silver Spring, Maryland, Mark-Amtrak collision in February of 1996, a very serious event, indeed.

The concern, if one can infer it, seems to have been that there were three crew members, all three crew members were in the small control compartment of the Mark train immediately prior to the accident, and I think there was some interest on the part of investigators as to what in the world these folks were talking about that could have caused the engineer to have forgotten or to have disregarded the signal indication that was provided immediately prior to the station stop.

We've already taken action under Emergency Order Number 20 to deal with the immediate issue there, and we're working toward fulfillment of Board recommendations related to positive train separation and positive train control, but we do have this voice recording recommendation apparently driven to some extent out of that accident, although I understand there are other accidents where this is certainly of interest.

Obviously there's a public need here to have this information. Misunderstandings can occur in the cab, and they do occur. If it's between crew members rather than between crew members and the dispatcher, there will be no record of that currently. It's certainly possible for crew members to be in a state of inattention, failing to call signals, for instance, or to be distracted as may have been the case in events such as the Silver Spring accident.

What are some of the other sides of this from the railroad industry perspective? As we've heard it from our customers, first of all, we do have dispatcher tapes for that leg of the communications chair. The data recording that we have on board does monitor bottom line performance in an increasingly-capable manner, and a lot of the interest that would swirl around the issue of cab voice recording would have to do with keeping trains apart, and we have at least a concept in positive train separation that will address this need; that is, the countermeasure is already identified.

There's a question as to when PTS is going to be affordable and deployable and so forth, and the Railroad Safety Advisory Committee is working on that, and we'll have a report to add to the literature on the subject, and we have active projects going on to develop and deploy the technology.

But that is certainly some of the information that we get back from the regulated community.

The remaining value of cab voice recording would deal with crew interaction, distraction of crew members perhaps. Some of that can be inferred from circumstances, and one of the arguments that's a fairly persuasive argument in terms of the history, I guess, of cockpit voice recording is in most cases in the railroad industry, we have surviving crew members who, if they can be persuaded in an incentivized system to be forthcoming, may be capable of helping us understand how these human failures occur.

So, it may be arguable at least in terms of the original purpose of cockpit voice recording that the aviation paradigm really doesn't apply here.

There's a down side to listening to folks all the time. Workers in the industry disfavor this approach. There may be a temptation, if the technology's not sufficiently secure, to use the information for collateral purposes. Obviously that could be taken care of.

On the other hand, there would be a price to pay in terms of the administration of any such program because we, unfortunately, even in the midst of extensive culture change in the railroad industry, and it is going on, I think it really is happening, we still have very discernible pockets of paranoia out there, folks who reject at all costs reasonable explanations because of the unfortunate circumstances that they've been in in the past, and there would be a public cost in terms of administration of any rule that called for this voice recording approach.

Right now, we're wrestling with all this stuff. The particular problem that holds us up at the moment, it seems to be a showstopper, is that currently in the railroad industry, there are no effective statutory barriers regarding misuse of voice recordings after an accident.

Statutory provisions do exist for aviation specifically by the National Transportation Safety Board which is the repository of the data and the analyzer of the data, and we simply don't have that kind of statutory protection at this point. The potential for misuse is substantial, and it's probably most substantial in situations where the information to be gleaned is least meaningful, and that is in events involving highway rail grade crossing accidents and trespasser fatalities, where the temptation to use extraneous conversation in the cab is a basis for a liability claim might be too tempting for some in my former profession, I guess I'm still in the profession, to avoid.

If we had statutory restrictions, they should apply to railroads, although we probably could handle this prong of it through regulation, as I believe the FAA has done.

We certainly would need restrictions on the Board and FRA with respect to inappropriate publication of the information, and also on the courts, as is in the case in aviation, to avoid inappropriate discovery, and that generally requires that a judge examine the data in camera, which is behind closed doors ex parte (without the other folks there).

So, where does that take us? We're not for positive sure. We're going to undoubtedly be wrestling with this for some time. My friend Barry Sweedler was lobbying me out in the foyer there, giving me more to think about, and I think that as we attend forums such as this, inevitably we learn more about what the potential may be, and also what some of the problems may be with respect to voice recording.

We do know that we need to use the data we have more effectively. Event recorder data today is abundant and potentially very useful for a variety of preventive and remedial purposes. The integration of locomotive cab electronics is opening a variety of new opportunities to get data on a cost-effective basis, and we're exploiting that in the Railroad Safety Advisory Committee Working Group.

It's clear that as we mature what we call the safety assurance and compliance program, which is Administrator Molitoris's approach to the issue of regulatory compliance and enforcement and safety enhancements on the railroad, which we deal in a partnership basis with labor, management, try to get safety improvements, that we need to be looking at non-fatal post-incident evaluations, and we need to be finding ways to immunize people and to get them to talk about it, and we also need to be talking about more actively and openly the relationships among crew members in the cab as we glean that information, and I know that that is not a new subject in other venues, but undoubtedly we need to be open to suggestions in the area of crew resource management.

Obviously we need to press for the implementation of positive train control systems. One of the difficulties we have here is we have so many things we want the industry to do that every time we add something to the list, we distract the industry from some of the unfulfilled items on the list, and then we're going to have to further evaluate our remaining data needs, and undoubtedly be in the conversation about voice recording and video recording and so forth for a good long time.

Thank you very much.