Mobile Mapping Underground: They Said it Couldn’t be Done

By Maraliese Beveridge

Built to connect a network of surface track and underground tunnels, a railway system on the east coast was tasked with the systemwide installation of cellular service to provide increased safety and connectivity. While cellular service had already been available in all the train stations, there was still no continuous service available inside the tunnels, some of which were as much as 26 miles long underground.

Constructed by assorted contractors, this unique system spans over 100 miles, has nearly 100 stations and provides service for more than 600,000 customers a day.

Converging as a cohesive entity, the track weaves back and forth, over and under, three states. The tunnels, which were built at various points in time, were hewn through rock and other materials on different terrains creating an inconsistency in the exact configuration of tunnel interiors. 

In order to design and build a modern communication system, a highly accurate as-built map of the entire system had to be created. Once completed, a series of cables could be attached to the tunnel walls that would provide the needed cellular service. Because the tunnel’s dynamic envelope has such limited space within, this map had to be of the highest quality available. Today we have the technology to do this and it’s called LiDAR. Light Detection and Ranging is a technology that uses sensors to scan a subject by sending out a physical pulse of light that measures distance based on the time of flight between when the pulse leaves the sensor and when it is reflected back and returns. This high-definition technology is so dense that it can measure over a million points per second per sensor. The term Mobile LiDAR refers to mounting high-accuracy mobile mapping scanners, in this case, to the top of a hi-rail truck that traverses the railway and scans while in motion. Mobile LiDAR depends on the use of Global Positioning (or GPS) technology for positional accuracies. Given that the tunnels were underground where GPS signals are not present, another method was needed. The team at Colliers Engineering & Design worked closely with the authority to develop a way to trick the GPS. So, how did they do it? 

The Team

First, the railway authority had to find a team of professionals who were up for this challenge and approached Colliers Engineering & Design, at the time Maser Consulting–on their conviction that it could be done. Not only did the team have the right combination of equipment, after a great deal of discussion the authority was convinced they had a viable approach and qualified team, so they championed the innovation. 

“No one had ever performed mobile mapping underground for more than a few miles,” said Paul DiGiacobbe, Director of Geospatial Services at Colliers Engineering & Design. “We knew it would be difficult, given the challenging geometry and sheer length of the tunnels.” 

The project team, consisting of three full-time survey crews, the Mobile LiDAR team, and key members from the authority, performed a 10-mile pilot study in which the initial results were validated and further substantiated the team’s confidence in the project’s success.

“The pilot study was key because it gave us the opportunity to learn how to deal with the performance of the Mobile LiDAR system without GPS,” explained Clay Wygant, Mobile LiDAR Manager at Colliers Engineering & Design. “We captured enough data to test and confirm our tunnel mapping theory. This set us up for success for the remainder of the tunnel system.”

Easier Said than Done

The design and installation of cell service had to be accomplished by fitting pre-fabricated radiant coaxial cable onto the tunnel walls using a cable tray throughout the entire 104-mile tunnel system. It was also required to have a safety clearance verification performed which ensured that the new cabling system was far enough away from the passing train cars. 

Radiant, or leaky, coaxial cable is a type of cable designed to both transmit and receive wireless signals and can be utilized by multiple carriers. Aside from being extremely expensive, the length of each cable was critical and had to be measured with laser-accuracy as it had to be segmented in predetermined lengths before installation. Each cable run also had to accommodate an installation path that wove from the inside of one curve to the outside of another across the tunnel while circumventing other appurtenances inside the dynamic envelope. So, every measurement counted and the only way to achieve that kind of accuracy was by using the 3D solution that LiDAR brings to the table.

Hence the Challenge

Rail maintenance personnel work in an unforgiving environment. The number one rule of trains is that the schedule never stops. So, to get the job done, portions of track were only shut down for short periods of time overnight during non-revenue hours to allow multiple survey teams to get their work done. 

In order for a hi-rail truck carrying the LiDAR instrumentation to physically enter the track zone, it has to be mounted on the rails at a rail/roadway crossing or in a rail yard. The closest crossing might be miles from the job site, so that factor alone can eat up precious track time.

So how was it done?

To accomplish the mobile mapping of the 104-mile tunnel system, the project team established survey controls from the surface station traversing down into underground stations. Then, scan targets were placed and surveyed along the tunnel corridors. This method continued to the next station where the team would traverse back up to ground surface checking into their GPS control.  

There were challenges presented by various coordinate systems historically used. Combining, translating, matching, and checking was an essential task in the making of this continuous mapping network. The authority developed their own low-distortion projection coordinate system designed to minimize mapping distortion associated with curved-to-flat mapping surfaces however all as-built data had not been re-projected at the time of survey. The approach was to traverse the entire tunnel system with the idea of capturing multiple survey datums and bring them into one. That way, they could use this datum through all the jurisdictions and serve-up multiple data sets and convert it on the fly.

With a control network established, the team would rely heavily on a key component of the Mobile LiDAR system: Inertial Measurement Unit (IMU). The IMU accounts for movement of the sensor system as it travels through space, adjusting for the smallest deviations and correcting the position 100 times per second.

“The difference with our team was confidence and a mindset.” Wygant explained, “Even though we knew exactly what the LiDAR equipment’s capabilities were, and how it was technically built and supposed to function, we had to compensate for the lack of GPS. We knew that the Inertial Measurement Unit (IMU) should perform as we anticipated and were initially confident yet concerned that when we reviewed the data it would not be looking upright with the track always down. We needed to see that it wouldn’t look like a corkscrew.”

The IMU did perform as expected and was key to realizing their success.

The software they used was also not built to manage such a wide variance between actual and captured location. The positional error (horizontal and vertical drift as well as a longitudinal expansion and contraction) was well beyond any pre-defined error index in the software. But that was expected. They trusted the integration and robust qualities of the technology, and their experience provided the confidence they needed to know that it could be accomplished, successfully.  

“Colliers’ approach to project management also enabled us to have the latitude to be inventive, creative and to take risks,” Wygant explained. “The existing geometry of the tracks was a known quantity derived from as-built plans which provided a relatively accurate guide as to how the track was constructed. The plans showed the geometry of the curves, horizontal and vertical position and superelevation of the alignment. We used this valuable information to supplant the GPS positional information that we were lacking in the tunnel environment.”

The team used several different criteria to adjust the Mobile LiDAR model to what they knew about the control that was measured by distances and defined a hierarchy of controlling data.  This ranged from highest (traversed and direct measurement) to known objects seen in scan and measured by various survey methods (this included pull boxes to rail components) to ancillary information gleaned from records or as-build documents to common structures and features seen by both opposing track runs in dual-track tunnels. Through the multiple iterations of post-processing the datum, redefining and quality control, the team arrived at a solution they could stand behind. 

After collecting the LiDAR mapping, the team routed the 3D position of the cable tray system based on the LiDAR mapping results and intended locations. They checked the dynamic car body to ensure clearance to the proposed cable tray locations.

Accuracy 

At the conclusion of the project, their relative accuracy ended up being within hundredths of a foot making it extremely accurate for clearance measurements. Their absolute accuracy, which is their placement on planet Earth, was dialed-in to within one foot. This was necessary to make the factory cuts for the leaky coaxial cables fit the field conditions and was a huge accomplishment. 

“All things being equal, all Mobile LiDAR post-processed to project control should match the provided values at control. It’s what happens in between control that is in question,” Wygant said. “Our extensive review, ever expanding site knowledge and dedicated champions all contributed to the successful creation of base mapping data spanning the entire tunnel network.”

Conclusion

Once the complete railway and tunnels were scanned, it produced both an accurate condition survey as well as the most accurate as-built the rail authority has ever had. Perhaps the bigger achievement was getting this information into people’s hands. Getting it to the rail authority in an accurate, digital manner that would help guide their future was invaluable.

The successful upgrade of this system mapping was accomplished through a combination of technology and human-power. As a result of the combined efforts and innovation on this project, Colliers Engineering & Design has been awarded a US patent for Tunnel Mapping. The patent recognizes the utility of Mobile LiDAR in GPS-denied environments combined with alternative means of geometric trajectory methodologies. This paves the way for other transit and rail tunnel systems to be mapped with high accuracy and repeatability.

And they said it couldn’t be done.

Maraliese Beveridge is Senior Technical Writer at Colliers Engineering & Design.