Magnetic climbing robots for safer infrastructure inspections

Bridges are inspected for safety by human operators using a combination of visual assessments, hands-on evaluations, and specialized tools. Operators often erect scaffolding or use cranes to access the structure, which can be costly in terms of time, effort, and logistics. Advancements in robotics offer a promising tool for efficiently assessing the health of Pennsylvania’s infrastructure while minimizing risks to human operators.

Researchers at Carnegie Mellon University have partnered with HEBI Robotics, a Pittsburgh-based modular robotics company, to develop a new class of climbing robots capable of scaling vertical surfaces, transitioning between connected areas, and overcoming small obstacles—all while carrying an inspection payload. These robots, also known as crawlers, enable inspectors to evaluate structural conditions with high precision, eliminating the need for scaffolding, crane trucks, or rappelling to take measurements.

“We're interested in building robots that can climb up infrastructure and overcome obstacles to conduct inspection tasks,” says Aaron Johnson, associate professor of mechanical engineering at Carnegie Mellon University. “An operator can be safely on the ground at the site of inspection. They wouldn't need any heavy equipment, like cranes or cherry pickers, and can keep a safe distance from dangerous situations or potentially hazardous materials.”

Most existing magnetic climbing robots for infrastructure inspection rely on large, fixed magnets beneath the robot. These crawlers must stay close to the inspection surface to maintain sufficient magnetic adhesion, making it difficult for them to transition between surfaces or navigate obstacles. Partnering with HEBI, Johnson’s team explored using magnetic wheels for adhesion, allowing the robots to climb and navigate steel structures more effectively.

“What sets our design apart from other crawlers is its ability to transition between surfaces and overcome small obstacles,” says Johnson. “With the wheels themselves serving as magnets, the crawler can stick to the ceiling of a structure while keeping its back wheels on a connecting wall, making a full transition to the ceiling possible. It’s capable of navigating from the floor to the wall to a ceiling or overhang without falling.”

The crawler's magnetic wheels also improve its grip on structural surfaces, allowing it to carry heavier payloads used to assess the integrity of bridges and other steel structures. Inspectors often rely on portable X-ray fluorescence (PXRF) spectrometers to perform elemental analysis, identify surface contaminants, and detect structural weaknesses. These specialized sensors can weigh over a kilogram—more than most crawlers are designed to carry.

“If you think there is mercury, lead, or other potential health concerns on the surface on an inspection site, the human operator needs appropriate personal protective equipment (PPE),” says Johnson. “With our crawler, we're not worried about lead exposure to the robot. The robot will be okay, and the human operator can stay at a safe distance without needing PPE and a respirator.”

The team believes that their crawler can obtain better quality data than other climbing robots due to its ability to collect measurements while in motion. Johnson says that if an operator wants to move the sensor along the surface at a controlled rate, the robot will be more efficient at collecting data than a human.

“As we collect measurements while the crawler is in motion, we can intelligently choose the next spot or location for another measurement based on the data we’ve collected so far, rather than just having to have a uniform reading—say, every three or four meters, for example. Using the PXRF, we can look at the data we’ve collected in real time and perhaps say, ‘The readings for weak spots or contaminants are getting a little higher. Let’s take some more readings in this area.’ Conversely, if you're in an area where the readings for contaminants are very low, then you can choose to spread out the readings a bit more.”

The end result is a more accurate estimate of the average distribution over a given area in the same amount of time.

“We've demonstrated that this process reduces your error rate,” says Johnson. "If you have a fixed amount of time or a fixed number of samples to collect, you can get a better average over that area by using this kind of wheeled platform—rather than a drone or another option.”

Johnson explains that, given a fixed time budget or accuracy requirement, operators can quickly deploy the climbing robot to collect data safely and efficiently. The total time required would be significantly shorter, since traditional inspection setups involving cranes and scaffolding are unnecessary. As a result, safety is also improved, with operators remaining on the ground rather than working at height.

The research team is optimistic about the project’s potential to improve infrastructure inspections, increase efficiency, and protect human operators.

“The exciting part for our research team is the access that our crawler robot provides,” says Johnson. “The mobility and the way the robot attaches to the structure to be inspected means that you’re not limited to flat surfaces only. Our design is capable of going over rivets or seams and transitioning from one surface to another, which is more of a challenge for other crawlers. With our crawler, you're able to get more complete picture much more efficiently. You can conduct inspections at lower costs with greater frequency and get more data out of each inspection.”