Spot Inspects Confined Spaces for Caltrans

The first time Ken Sutterfield used Boston Dynamics’ mobile robot, Spot, in the field, he estimates the robot saved the California Department of Transportation (Caltrans) upwards of $60,000 in less than ten minutes of work. 

Based in Eureka, Calif., Sutterfield is a technology coordinator in Caltrans’ construction division. In 2025, a culvert (or large pipe) running deep beneath a highway in northern California split apart. This caused soil to shift downward, creating a sinkhole in the roadway above. To complicate matters, engineers had to contend with a steep slope to one side of the highway, or risk causing a landslide.

“We were concerned that when we cut the culvert out, we were going to destabilize that slope,” said Sutterfield. “The other issue was the culvert was fairly deep. Once we get past 20 feet of depth, we have to have custom engineered drawings for that specific site. So we didn’t know whether we were going to affect that side slope or how deep the culvert was all the way through.”

The Dangers of Confined Spaces

State departments of transportation routinely work in enclosed spaces that aren’t designed for humans, but still require periodic inspections. These include stormwater culverts, siphons, utility vaults, pumping stations, and plenums above tunnels that house ventilation equipment. 

Many of these structures contain critical utilities—such as electrical conduit, fiber-optic cables, and natural gas lines—that crews must repair or relocate during construction. Depending on the circumstances, safety regulations can include atmospheric monitoring, ventilation equipment, and even traffic shutdowns. Such requirements exist for good reason, but they add significant time and cost to even routine inspections.

“Safety is a big thing at Caltrans,” said senior transportation engineer Aaron Chamberlin, who oversees field technology investments for Caltran’s construction division. “We have 1200 active projects right now for about 14 billion dollars in construction. Whether it’s a multi-billion dollar bridge or a tunnel project, if it’s on the state highway network, we’re dealing with it.”

“Confined spaces, damage assessment, and landslides were what we were originally thinking for Spot. A good rule of thumb is that any time you can use one of these robots in a confined space, it’s going to save you money.”

Aaron Chamberlin, senior transportation engineer

“Want to Try a Robot Dog?”

Chamberlin decided to purchase Spot for his department after learning that a colleague at the Alaska DOT was using one to keep wildlife off of airport runways. He sent Sutterfield a simple text that read, “Hey, do you want to try a robot dog?” 

Chamberlin figured if Spot could handle sub-zero Alaska temperatures, it could manage just fine in Northern California weather. He envisioned using Spot to carry cameras and sensors into dangerous spaces that would otherwise require time-consuming safety procedures. 

“Confined spaces, damage assessment, and landslides were what we were originally thinking for Spot,” said Chamberlin. “A good rule of thumb is that any time you can use one of these robots in a confined space, it’s going to save you money.”

3D Scanning in Motion

The culvert repair was Sutterfield’s first opportunity to put the robot to work in the field. (Sutterfield calls his robot “Stompy” due its marching style of walking). Using a wireless, multitouch tablet, Sutterfield drove Spot down a steep embankment and positioned the robot at the mouth of the 300 foot long, 42-inch pipe running under the highway. 

Mounted on Spot’s back were several key sensors. In addition to a 360° PTZ camera, a long-range wireless radio, and bright LED lights, Spot carried the main attraction: an Emesent Hovermap Simultaneous Localization and Mapping (SLAM) lidar unit. This is a mobile 3D mapping system that uses laser pulses to generate a point cloud image representing the exact geometry of a given interior. 

Sutterfield has ample experience using lidar, but Spot introduced a new way to utilize it. Traditional survey teams would need several static scans at fixed intervals—each requiring setup time and line-of-sight conditions that are almost impossible inside curved culverts. Spot made it possible to scan the entire interior of the culvert in a single pass. 

“We already knew what we would collect with the SLAM lidar,” said Sutterfield. “We needed a new platform to safely carry this system into different types of situations.”

Problem Found in Minutes

Sutterfield drove Spot over loose gravel, tree limbs, and mounds of dirt with ease. Viewing the live camera feed, he located a two-foot fracture in the pipe. Crews on the roadway above only had to stop traffic for five minutes during the actual lidar scan. 

In less than 10 minutes of operation, Spot had collected  a complete, geo-referenced 3D image that revealed the curvature of the pipe, the joint that had separated, and even traffic cones that had fallen from the roadway above. By that evening crews had what they needed to determine the trench size needed to expose the culvert, while still preserving the stability of the slope. 

“This let us hone in on exactly where the failures were on this particular culvert, and they directly correlated with the sinkhole above,” said Sutterfield. “Then I reviewed how it affected the side slopes. It cut right next to the edge, but not into that slope.”

Immediate ROI

Chamberlin estimates Spot helped crews avoid $50-60K in permits, unnecessary excavation work, and potential environmental mitigation. The robot’s ability to walk over rough, muddy terrain only further impressed Sutterfield. 

“I’ve been surprised at how well the robot can keep itself upright in even eight to ten-inch deep mud,” he said. “I really didn’t expect him to be able to do that.”

Inspecting Bridge Cavities

Culverts aren’t the only confined spaces where Sutterfield has deployed Spot. A second major deployment involved inspecting utility lines inside a hollow portion of a concrete bridge called a soffit. The utilities needed removal, but in order to properly determine a budget, Caltrans first wanted to inspect the access within the soffit.

These soffits are created when construction crews pour concrete around wooden framing that gives the bridges their shape. Engineers typically leave the wood inside, as it’s too expensive to retrieve it. Unfortunately, trespassers will crawl inside and strip out the framing, blocking or damaging utility lines that may be inside. Sutterfield sent Spot in to take a look. 

Spot’s images confirmed that vandals had indeed piled wood directly on a utility line. The obstruction meant the contractor wouldn’t be able to follow the original plans. This helped Caltrans avoid an emergency change order and ensure the work could proceed with necessary safety precautions.

“We were able to determine the utility was blocked,” said Sutterfield. “That allowed us to include the contingency funding up front instead of finding out later in the middle of construction.”

Expanding Spot’s Role

Sutterfield sees future uses for Spot in other areas, especially tunnels. For example, plenums are spaces above tunnels that use exhaust fans to expel vehicle exhaust and smoke. They also typically contain utility lines. During inspections, the exhaust fans have to be turned off to protect employees, which also means traffic must be stopped to avoid exhaust building up. However, Spot will be able to conduct inspections without impacting traffic. 

“All of that equipment, plus the concrete structure that keeps that plenum open, needs to be inspected,” said Sutterfield. “Those are all opportunities for robotic inspection without having to close down the highway or the exhaust fans to review that stretch of the tunnel.” 

Early Results Promising

Interest in Spot has expanded beyond the division of construction, as well. Maintenance crews, culvert inspectors, and structural engineers are now evaluating how robotic inspections could support their work. However, these teams are just beginning to explore what SLAM lidar can do for them.

“Most crews don’t yet know what the sensors can do,” said Sutterfield. “Before thinking about the robot, they need to understand the data.”

Nonetheless, as far as the construction division is concerned, Chamberlin believes the early results justify continued exploration. He has already purchased a second Spot for the southern portion of the state.

“Two robots might be enough, or we may justify more,” said Chamberlin. “It depends on how often the use cases show up. Anywhere we have difficult locations to access—culverts, bridge cells, landslides—if the robot keeps people out of harm’s way, that’s the value right there,” he said.