Sensors and Systems
Breaking News
CHC Navigation Introduces the i73+ Pocket-sized GNSS Base and Rover with Built-in UHF Modem
Rating12345The i73+ IMU-RTK is the most compact, lightweight and...
RMSI Appoints Arun Vishwanathan as Vice President – Business Development
Rating12345Dallas, Texas: RMSI has expanded its global sales team...
Kavel 10 Aerial Mapping pioneers from the Netherlands chose Phase One PAS 880 system to improve productivity
Rating12345Best in Class Image Quality with Outstanding Time-Saving Workflow...

January 24th, 2011
Positive Train Control Elevates Role of GIS at CSX

  • Rating12345

ErikHenderson_thumbThe mandate by the Federal Railroad Administration in the United States for Positive Train Control (PTC) places a whole new requirement on the commercial rail providers for cooperation and communication of train movements to ensure safety and security. The level of precision in the mandate pushes the technological capabilities, while also offering whole new levels of efficiency. V1 editor Matt Ball spoke with Erik Henderson, GIS business consultant at CSX (one of the country’s largest railways), about the technical requirements and business process implications of this mandate. Henderson spoke at the Trimble Dimensions conference about the field data collection requirements and planned solutions, and conducted the interview immediately following the presentation.


The mandate by the Federal Railroad Administration in the United States for Positive Train Control (PTC) places a whole new requirement on the commercial rail providers for cooperation and communication of train movements to ensure safety and security. The level of precision in the mandate pushes the technological capabilities, while also offering whole new levels of efficiency. V1 editor Matt Ball spoke with Erik Henderson, GIS business consultant at CSX (one of the country’s largest railways), about the technical requirements and business process implications of this mandate. Henderson spoke at the Trimble Dimensions conference about the field data collection requirements and planned solutions, and conducted the interview immediately following the presentation.

ErikHendersonV1: One of the things that you touched on in your talk is the baseline accuracy for purpose of the different vertical asset classes (tracks, signals, etc.). Did each department have different accuracy requirements before the Federal Railroad Administration mandate of Positive Train Control?

Henderson: Before the mandate came, each department was responsible for an asset. Train operations uses signals, but our signal department maintains the signal. Both departments had a need, and both maintained a separate database of signals. Network operations isn’t as concerned about the accuracy. The conductor of the train knows there’s a signal five miles ahead, but not plus or minus a hundred yards. They’re driving a train that needs to stop before the signal, and don’t really need to know accurate position.

The signal department knows about the location and the length and gauge of the cable that connects them underground. For them, it’s really critical that they know all about their assets, including the location of the asset. In tomorrow’s world, after PTC the horizontal positional accuracy needs to be at 1.3 meters or less at all times.

V1: How will you go about maintaining this level of accuracy?

Henderson: I’ve met resistance internally regarding the communication of assets that have been moved, and the fact that individual departments can’t move anything without letting us know its new position. They ask about the accuracy of 1.3 meters, and some have questioned whether they need to let us know if it moves less than that amount. The problem is if someone moves it less than 1.3 meters, and then next month someone else moves it less than 1.3 meters, then it becomes an issue where we’re no longer accurate.

We’ve had to put some rules in place to be notified if anything moves greater than a foot, and will need to go out and re-survey anything that moves. We’re talking about a network of 21,000 miles of track and tens of thousands of assets.

GIS will be the spatial system of record for all of these assets. We are going to build linking tables (XREF) to historical systems of record, such as the signal database with unique IDs that point to the spatial record. Tying in the signal records to the spatial record is a painful and tedious step that looks at each record one-by-one and ties those in. But what that gets us is that the GIS system can move assets without having to worry about duplicating that change in all the separate legacy databases.

We will maintain each record twofold. The spatial representation will be in the GIS, and any asset information that the individual departments want to maintain can continue to be stored in their own database.

V1: Before the mandate was there a different positional accuracy for assets for the different business units?

Henderson: Traditionally everything was mapped to the mile post. If you were in one of our divisions and wanted to make a map of your data, you would give us your assets with a mile post associated with it. We have a linear reference system (LRS) with centerline, track and mileposts. Whatever you have plotted to the milepost, we can plot it in our LRS and produce a map, and we typically received annual updates that we would map. Now the engineering mile post is a data layer, and we have direct connections between the various data sets.

V1: Did this rough mile post accuracy limit the utility that you had from the system in terms of analysis of the network and other business processes?

Henderson: We ran into issues with network operations, where they had a software operation that limited them to 10 assets per mile represented in their on-board display. If you had a set of mileposts in Chicago, and there were 15 assets within that mile, you took the additional five assets and bumped them outside that mile.

Back to train operations, the train cares what its next set of targets are, and the next thing that’s coming up. Engineers are trained to know their territory, so they know how far the next signal or switch is. In a database they don’t care whether its after or before mile post one, so they would manipulate the mile posts accuracy to fit the software constraints. In tomorrow’s world we can’t do that. We’re having to force that software vendor to handle more assets per mile.

The accuracy is now going to be absolute. To our benefit, we have an FRA mandate that says when I survey assets that’s what goes, and that’s its final location.

V1: In the mapping world, we understand a range of positional accuracy. Are you faced with educating about accuracy?

Henderson: Historically, if people wanted to be cutting edge in the railroad industry they would take out their fish finders or even cell phones and record the Lat/Long. I received a database about three months ago with our communications and signals offices, and these used to be two separate departments, and five years ago they merged. The director asked the forty or fifty offices to go to Google Earth and write down the Lat/Long.

They supplied me with a spreadsheet and asked for a map because they were thinking of consolidating some offices. It was a great little project. With 45 people around the country each transferring their Lat/Long, it was a really interesting spreadsheet. It was a great amount of effort but, some were state plane coordinate, some were decimal degrees, it was everything under the sun. We had to go clean up the data.

The level of accuracy was based on the need, because I don’t need survey grade for an office location. But on the flip side, we have engineering folks that were designing new capacity, such as new sidings and main tracks, and we need engineering-grade accuracy.

John Chance Land Surveyors is flying our whole network right now as a one-time snapshot and their giving us 4 centimeter-level accuracy. Ultimately, I want to be able to record everything at our original recorded accuracy. I’d like everybody to have to meet 4 centimeters, but the volume of change that we expect doesn’t meet that level of accuracy, so we’re trying to find that happy medium.

We’re back to the, “if it moves a foot, let me know.” If I said, “if it moves a centimeter,” I don’t know. Do railroad assets really tend to move a centimeter? You’re not going to get a signal mast that will move a centimeter, so it’s not realistic. We’re trying to put some realistic properties around how accurate we’re going to keep things. We made a significant investment to fly every subdivision of every corridor of the network, so we want to keep that up to date.

V1: It’s the mandate that is pushing the accuracy requirement, but there’s a benefit to the utility of the data that is being collected, right?

Henderson: Now we’re the system of record, and there’s no argument in regard to the position of the asset. We also have some things that we’re able to say based on the mandate that eliminate arguments about how we collect and represent the information and how the database is built. Now we’re going to know exactly how many assets we have and where they are.

We operate in 23 states and we have 10 divisions, with a typical division covering two states. Now if we look at an entire division and map the number of signals in that division, we know the exact number of signals to a high degree of accuracy and confidence.

A lot of the departments have decades-old databases that started with pencil and paper. They understand that we’re very accurate and that we’re visiting every location both by helicopter and by high rail with aerial photography and LiDAR. We have forward facing and down facing cameras for digital stills and digital video. The pixel resolution is fantastic at 4” resolution. We can see the cross ties, and read signals and signs. Even the lasers are returning RGB values for colorized laser point clouds. We started flying early this year, and will be done some time in 2012.

We’re also going and ground verifying all of our assets, because there are some things that the sensors can’t pick up, like the DOT number on a road crossing. That’s a sign that typically faces drivers, so the helicopter can’t read it. We’re driving the high rail to record speed signs and mile posts and other ground attribution that you can’t get from aerial. We’re augmenting the databases of each department, but they’ll still come back and cross reference our spatial location.

The helicopter LiDAR data will always be the most accurate location, but if someone on the ground says that something is a mile post not a speed sign, we’ll defer to the person on the ground on that judgment, but we’ll give it the LatLong from the helicopter because it’s the most precise spatial measurement.

A lot of these departments are starting with a fresh inventory. It’s all part of the reconciliation process of building XRef tables to relate to our record of positional accuracy. In the signal department, they have to give us a 1 for 1 record with metadata on each signal, who built it, who maintains it, what are the electronics inside, what it controls, all that business information that they have to associate with one of the signals that we surveyed. They are going through, asset by asset, and matching it up to one that we surveyed and making the unique cross reference table.

V1: What are some of the implications of having all of this information together?

Henderson: One of the big benefits that we’ve talked about relates to the Nashville floods where we had quite a bit of track that was washed out. We had previously flown it, so they came straight to us and asked for an exact asset count and configuration. We were able to provide the exact amount of track, the number and type of switches and signals, and all of the other assets.

For every mile of track, we know exactly the costs, and we new exactly what the damages were, and we documented it. It’s much more realistically auditable, where our old databases were much more hit and miss. We can now go back and redesign.

We’ve had people talk about Katrina similarly, with a backward ROI, realizing what we would have been able to produce. In hindsight, it’s becoming a good thing, because we can say, just think if we had this.

V1: With the high degree of track knowledge and exact understanding of grade changes, won’t you be able to save fuel and maybe automate the engine more?

Henderson: We have an entire team that is working on fuel efficiency, because there is huge savings to be had in fuel. Right now it’s the engineer’s job to get that train from A to B, no matter what it takes. If the train is too long or if the grade is too steep for the weight of the load, they might get the biggest locomotive and put it on the tail end of the train and push to get the job done.

Now, we are able to calculate the grade, the size and weight of the train, and the class of locomotive needed to move it safely and efficiently. What they are going to be able to do is to create the “golden run” so that any time you go from Birmingham to Nashville or from West Virginia to Pennsylvania, we know what the slope and distance are. The slope won’t change, but the size of train and the conditions, including weather, will change. So when we put the factors in, knowing the track to a very accurate level, then we’ll be able to dictate the locomotive power required.

We’re also now tracking the real-time operation of the train. If the engineer  goes full throttle too early, we’ll be able to tell him later on that he should wait until a certain point.   Similarly, we can tell them to start braking before the crest of the hill if needed, not waiting until the downgrade to slow down, and not having to go full brake. The locomotive team is studying this sort of efficiency issue, and are teaching our engineers how to be more efficient, which translates into huge fuel and cost savings.

V1: Is there any fear from the workers that you’ll just fully automate the engineer’s job?

Henderson: Out of 10 questions that I get in every meeting, at least three or four of them are related to staffing. They ask if we’re going to a one-man crew, or removing the signal system, or remotely controlling the trains.

We have remote control locomotives in the yards today. If you think about the airport trains or the Disney Monorail, they’re automated. We know exactly the length of the track and distances between points, and we’re just talking about scaling that up to a much larger area. We’re building the technology so that ten years from now it’s not inconceivable that we would be able to control them remotely.

The age of our workforce is increasing at CSX, and over the next 10 years we’ll be losing approximately 80% of our workforce to retirement and attrition. We also have a very experienced workforce, so over the next 10 years it will be critical to bring in newer technology and creating a better work environment so that we operate more efficiently. If you look at the history of railroads, we’ve gone from hundreds of thousands of workers to tens of thousand employees. We’re at about 35,000 people right now.

V1: The business case is certainly there, given the efficiency gains. Are there other ROI benefits other than fuel and potential manpower savings?

Henderson: I get people coming to me every day in the GIS department and asking about the surveying work that we’re doing. People from different departments want to know the assets that we have, because we haven’t had a company-wide means to look up this data. It has been a very tedious and time-consuming process for each department to do research on assets. We have to teach our internal clients about the mandate and about the benefits of the surveying work that we’re doing.

I’ve had some contact recently with our financial departments, such as accounts receivable and our tax department. Traditionally these departments haven’t been able to defend the cases brought upon them. If we charge “X” for a public crossing upgrade, but the states think they should only be charged “Y” based on the assets installed, we haven’t had the manpower to argue it, but we’ve had the resources to pay the bill. It was far cheaper to pay than to file a court case.

Electrical utility needs is another issue, because just like you and I, we pay our utility bill every month. We’re surveying all these assets, and we know that a signal is powered by electricity and a powered switch requires electricity. So, we’re able to build a database knowing there is an electric utility meter at each of the assets. We constantly sell off different portions of the track to different entities. We’ve been known to sell off a piece of our rail, and not tell every department in the company. If our accounts payable department keeps getting the bill and paying it, the little increments add up.

Now that we have very accurate asset records and reporting, our accounts payable department will be able to match assets to utility meters. Just starting now, we’re able to reconcile all of those databases. Knowing exactly the LatLong of every asset also gives us great tax benefit. We can overlay asset data with cities, counties and states, and with each having their own tax rates, we now know to a greater degree how we’re being taxed and able to discover if we’re being over taxed.

V1: The PTC mandate gives you just five years to complete, and it’s an ambitious project. Are you on target to meet the mandate?

Henderson: We’re a year into it now, and we have five years to go. We’re just coming out of the deer in the headlights stage now. There were so many departments that had to get together and discuss what it meant to us, and how the other railroads are doing it. We have to all be interoperable, so we can’t design our perfect solution without talking to the other railroads. We have to design a system that everyone can download each others databases in real time, so each of the databases has to be exactly the same.

We’re utilizing cutting edge technology and doing R&D around different solutions to see if they will work. Once we get some of these solutions in place that we talked about today, we can more safely ramp up manpower, and efficiently get more people working on it.

Each railroad is implementing a pilot territory, and once we get solutions working well there, we can start mass producing it, and spreading out along our network.

Leave a Reply

Your email address will not be published.