By making the post-construction survey the critical source of reliable asset information in the form of a 3D intelligent model, Ranvir (Ron) Singh, Chief of Surveys/Geometronics Manager, Oregon Department of Transportation, US, proposes to turn the road construction process on its head.
In a recent presentation you talked about “turning the construction process on its head.” Could you describe what you mean by that and what you think is the primary reason that your and other DoTs are interested in this?
The Oregon Department of Transportation (and I suspect we are similar to many DoTs) got into automating its engineering back in the 1980s. We began using CAD, road design systems, and digital surveying instruments. The end goal continued to be to produce hardcopy construction documents, but we wanted to speed up the whole process.
A few years ago, I wrote a paper saying that we have been working to achieve this goal for the last 25 years. It is time to call that initiative a success and move forward with a new goal. I proposed a new goal which would be all about data. We would create data not just to produce construction a digital model for the entire highway system’ ‘It would take generations to develop plans or for designing and showing the public what things look like, but create it throughout the entire lifecycle of a highway including construction, operation and maintenance. The concept of looking at the full lifecycle of a structure is not new. I think the concept is more prevalent in the vertical building industry than it is in highways.
The way we have been designing and building highways for a long time has been quite simple — we design a road, flatten it out on a piece of paper, produce a plan, stamp it by putting an engineer’s seal, and then build it. During construction, the surveyors read the plan and drive stakes into the ground to guide the equipment. The equipment operators have people on the ground who read the stakes and use rudimentary tools like pocket tapes and hand levels to guide bulldozers and other equipment to build the road. That is how we have been doing it. But all this has to change. At present, the contractor has to read the 2D plans and try to interpret our intent in order to build a given structure.
But a 3D model shows the contractor in photorealistic 3D what currently exists and that we want to build. More importantly, the 3D data can be used directly for automatic machine guidance. Today, we have machines that can automate almost all aspects of highway construction. For example, with a 3D design, we do not have to drive stakes into the ground showing where a culvert goes. We load the 3D data into the excavating machine. Either the operator will use the display he sees on a monitor in the cab and uses it to guide the machine or with certain devices like an automated curb machine, the system will position it to the correct line and grade and build the curb automatically.
What do you do about underground utilities in the current highway construction process?
Utilities are a huge problem. Depending on what we are doing, we need to know where those underground utilities are. Today, the utility companies help identify the utility infrastructure for us. They come and put paint marks on the road surface to show the location of the pipes and lines. Most of the time there is no depth information and the accuracy is typically plus or minus two feet [which is the level of accuracy that utilities are expected to deliver in Oregon]. The survey crews map those coloured paint lines in 2D and put them on their design to show where the utilities are. If the highway designer needs to know the location of underground utilities more accurately, we go out there with a backhoe, dig a hole and find the lines. The surveyor records the precise three dimensional position for that utility at that location. There are now technologies like ground penetrating radar (GPR) that allow you to locate underground utilities without digging, but from my vantage point it is still difficult to interpret and is very time-consuming.
Does the uncertainty about location of underground utilities introduce significant risk when designing a highway?
Absolutely it does. Today, if we have to build a new interchange on the Interstate Highway System, the typical process is to go and do a complete survey of the area. This provides fairly reliable information about what is on the surface because it records what is out there right now. But it does not provide information about what is underground. The question we have to ask is why didn’t we capture the actual conditions both on the surface and underground when the highway was first built. When we use remote sensing technologies like GPR, the wise thing to do is to show that utility centreline with a 3D buffer around it representing a cylinder or ellipse of uncertainty for that cable or pipe.
This sounds like creating a digital model of the highway during and after construction?
One thing I have been pushing for in my department is that we should do the biggest survey for a particular section of highway. I would like to see this happen during construction and post-construction rather than upfront before design. In this scenario, when we are building utilities or other structures underground, there are a number of tools that can be used to record a three-dimensional position for the pipes, cables or culverts. Then, as we build up the roadbed prism, we would be capturing all that information in 3D as the underground foundation is going in.
This way, when we complete the project, we would not only have a new physical road, but also a new 3D digital as built road. Over time, we would see more of these little pockets of 3D-engineered models that we could start using as the starting place for future designs. Years later, if we want to reconfigure this interchange, we do not have to do a full survey of the area again. We do not have to pull manhole lids to look inside and wonder what is in there. We will already know what is in there. We will already know the location of the pipes and cables, their materials, and everything we need to know about them.
But what if something changes in the intervening years, for example, the road is widened?
As an agency, we would have to have the discipline that when we change something through our maintenance activities, we either note in the database that something has changed and it needs to be surveyed or we get the surveyors to survey the change immediately after construction and update the database. We have to ensure business processes are in place to maintain the reliability of the digital database.
I recognise that it would take generations before we could develop a digital highway model for the whole system. As a first step, we are scanning out entire highway system with mobile LiDAR scanners and capturing the highway as it exists today. In addition to point clouds, we are also capturing imagery using a variety of reality capture technologies. Now somebody can look at a piece of highway on a computer system and measure in 3D. My goal is to cover our entire highway system with point clouds and oblique imagery, which can be overlaid with all the asset information that comes from a GIS system. You will be able to see all of our surface asset information very accurately. As we build new projects, post-construction surveys will help us generate more and more of these little pockets of reliable engineered information appearing within this big database. These could be queried, and since we would know that this came from a survey, we can rely on it for future engineering.
We are organising a two-day demonstration of some of these technologies on July 9-10 in Oregon on behalf of the Federal Highway Administration. We are inviting 18 western state DoTs to attend and we will show them 3D design and machine-control, including automated bulldozers, graders, excavators, paving machines, and curb machines.
Another long-term benefit of maintaining up-to-date 3D datasets would be the autonomous vehicles. Autonomous vehicles will be data hungry. Already cars are using real-time sensing of their location and what is around them. But I cannot imagine that they would not benefit from knowing about the highway infrastructure. For example, if an autonomous vehicle is driving down the road, it needs to know the roadway alignment and where certain things are — in case of an emergency can I cross this line and go in that direction or is there a cliff?
I believe that those vehicles cannot just sense that information and need to have information accessible to them in real-time. They also have to sense what is changing, where other things are on the road, cooperative objects like other smart cars and uncooperative objects like a cow that has wandered onto the road. We are talking about data that is not just used during construction, it is also used to operate and maintain our highway systems.