Image in modal.

Growing labor shortages continue to plague the construction industry, while trade schools and job training programs report their current enrollment won’t make up for a wave of retiring skilled workers. Robotic solutions to repetitive, labor-intensive tasks on construction sites are already available, but a new generation of robots aims to boost workforce efficiency by tackling the tedious and dangerous tasks that plague general and specialty contractors.

Drones and small walking robots such as Boston Dynamics’ Spot have been used by contractors for tasks such as reality capture for years now, but getting the most out of general-use robots on construction sites often requires the oversight of virtual design and construction teams and project managers, and many contractors simply don’t have the personnel or time to spare.

Not many people know how to take advantage of general-use robots such as Spot on the jobsite, says Conley Oster, co-founder and chief operations officer at Raise Robotics, maker of a robot that can install brackets for facade panels along slab edges. However, there is a “definitive use case” for robots in niche applications. “Contractors can digest exactly how they’re going to use [a robot], and then, as long as it’s easy to integrate into their existing workflows, it’s something they’ll adopt quickly.”

robotic arm

Raise Robotics offers a robotic arm that can do some of the more dangerous fastening tasks found on a jobsite, such as working on the edge of a floor plate.
Photo courtesy Raise Robotics

Construction robotics from firms such as Dusty Robotics, Canvas, Kewazo, Raise Robotics and Nextera Robotics are built around specific construction tasks, rethinking processes to make jobsites more efficient. Such purpose-built robots are already used by contractors including Gilbane Building Co., DPR Construction and RQ Construction.

“Improving safety [comes] first,” says Artem Kuchukov, co-founder and CEO of Kewazo, which developed a lifting robot to reduce the crew size needed to set up a building scaffold to only two workers. “The second thing is improving productivity on site. Especially when it comes to industrial sites, we see a lot of jobs being done with cranes. To replace those cranes is a big advantage.” Large cranes designed to move heavy-duty materials aren’t necessary for lighter weight materials, he adds.

Kewazo’s LIFTBOT, a mast-mounted materials transportation platform, can lift scaffolding segments to height as a scaffold is being assembled, as well as building materials along its set of rails.

Scaffold Resource used LIFTBOT on the facade reconstruction of the U.S. Capitol. The company’s Vice President, Bill Schikner, says the speed of material delivery using LIFTBOT, along with its ability to meet the Architect of the Capitol’s expectations for safety and quality, made it a good fit for the job. The robot itself is remote-controlled, and scaffolding guardrails don’t have to be moved or modified to deploy it.


ENR Construction Cost Data Dashboard
GET STARTED

“When we were thinking about starting up Kewazo, philosophically, we considered that construction is all about just two activities,” Kuchukov says. “One is how you put things together vertically, horizontally in an additive process, like Lego for adults. The second process is, how do you get this material where it needs to be? If you focus on getting material where it needs to be, you figure out how to deal with one type of material, and then you can quickly switch to another one.”

Dusty Robotics Field Printer 2
The Dusty Robotics Field Printer 2

The Dusty Robotics Field Printer 2 can print layout markings directly onto the floor slab, and is more maneuverable than its predecessor.
Photos courtesy Dusty Robotics

Repetition Drives Automation

Few tasks in construction are as labor-intensive and repetitive as spraying and finishing drywall. Wall panels can weigh 50-60 lb apiece and typically require two skilled laborers to install. Drywall mud air dries and can take hours to set up properly. Humidity and air quality can add to delays. To mitigate these issues, San Francisco-based Canvas has been automating drywall finishing since 2017. It recently released the 1200CX, the second version of its drywall robot, which is now more compact, lightweight and maneuverable than the original 1550 model.

The Canvas robots are “able to finish drywall to a Level 5 finish in significantly less time than even a very qualified drywall finisher,” says Levi Litke, self-perform leader at RQ Construction, based in Saint Johns, Fla. “Mainly it’s just because of the dry time between the coats. Our projects almost all require a Level 5 finish.” As the highest level, the drywall finish needs to be completely even across the wall surface so any applied paint will have an even tone. “It’s a lot of repetitive motion when you have big, flat walls. It takes skill, muscle [and] time working at heights” to achieve, explains Litke.

“Contractors can digest exactly how they’re going to use [a robot], and then, as long as it’s easy to integrate into their existing workflows, it’s something they’ll adopt quickly.”
Conley Oster, Co-founder and Chief Operations Officer at Raise Robotics

RQ has been using the 1550 on several of its projects across the country. Litke says the ongoing skilled labor shortage is the main reason it uses Canvas. When RQ has a chance to shorten its construction schedules, it needs to be able to take full advantage of those opportunities whenever possible.

At a project located at Guantanamo Bay, Cuba, the extremely hot and humid conditions extend drying times to more than a day. Typical human workflow requires three to four coats with a dry sand in between. With Canvas, Litke says “we’re able to do that with one to two coats. We’re shaving days and or weeks in different areas.”

Canvas CEO Kevin Albert says the company has learned a lot from the contractors and labor unions it has worked with about what they need as the company sends more and more robots into the field.

“We have robots in five different major regions right now all across the country, so we’re getting really good feedback,” he says, which led to improvements rolled out with 1200CX. At 30 in. by 34.5 in., weighing in at 1,200 lb, the robot boasts a 12-ft finishing height and is significantly more compact than the 1550, which was built on a mobile lift chassis and is able to work at heights up to 15.5 ft.

Where taller finishing heights aren’t required, such as multifamily residential and hospitals “where you have patient rooms that are 10 feet by 8 feet, [or] hallways where it’s hard to turn into a room … it was really the maneuverability of the 1550 that wasn’t sufficient,” Albert says. In addition to the smaller footprint, it features “all-wheel steering, so you can drive it in any direction at any time.”

Canvas is also seeing use of its 1550 in data centers and other high-tech projects with long wall sections. “I give huge credit to our customers who have continued to work with us and give us great feedback,” he says. “We’ve worked  to continue to incorporate every piece of feedback that we got into the machine and the results are incredible.”

The Kewazo LIFTBOT

The Kewazo LIFTBOT transports materials up a scaffold, getting them to the exact place they are needed.
Photo courtesy KEWAZO

The Little Layout Robot That Could

While not as labor-intensive as finishing drywall, multiple trades doing layouts on every floor of a building with chalk lines and pencils is a tedious task. Dusty Robotics has been looking to reduce the tedium since 2018 with the Field Printer, its BIM-focused, automated layout robot. Able to print directly and accurately onto a floor plate with minimum onsite supervision, the robot has seen extensive use in building construction. Now, a new model of their automated printer is taking the little robot from a novel approach to a go-to layout tool.

“People spend time on value-added problem solving, not data gathering, and smaller projects that don’t have the budget for robust documentation can get the same level of reports.”
Ryan Spotts, Gilbane Senior Project Executive

“We heard loud and clear from our current set of customers that they wanted something that’s smaller, more nimble, more agile and can fit into tighter spaces in order to get a more complete print,” says Tessa Lau, Dusty’s CEO. “That’s what we’re delivering with Field Printer 2. It’s directly driven from our experience. As of last year, we’ve printed over 125 million square feet of layout across tens of thousands of buildings. Based on that experience, we set out to design something that would solve some of the problems that we experienced on those earlier job sites.”

Lau says that conduit stub-ups all in a row often require layout markings between them, and that was one reason for a smaller form factor. The earlier model of the Field Printer required near-constant line-of-sight to a nearby total station to maintain accuracy and remain properly oriented, but cluttered jobsites create blind spots on the floor plate. This is not an issue for Field Printer 2, which can navigate by dead reckoning based on the layout data stored in its internal memory.

“In areas where it temporarily loses line-of-sight to the total station or the laser tracker, we can keep printing.” Lau explains. “That’s a big differentiator we have compared to our competition, and it gives our customers just a much more complete print.”

Dusty has also added new types of printer ink for text and other mark-ups that wouldn’t be chalked in during a traditional layout. Lau says customers also asked for smoother battery changes. Field Printer 2 uses off-the-shelf power tool lithium-ion batteries. More sensors also make it easier for the robot to detect obstacles and navigate around them.

site documentation robots

Gilbane has used Nextera’s site documentation robots to capture regular 360° images of project sites for AI analysis.
Photo courtesy Nextera

Getting on the Edge

Slab-edge work requires workers to tie in and out, remain constantly aware of potential hazards and accept the level of risk that comes with working at height. Raise Robotics’ robot is all about getting humans away from the slab edges of multistory projects. Early adopters include DPR, Lendlease, JR Butler Inc., Harmon Inc., and TSI Corporations.

“We match up nicely with manual output [cost-wise] but what you don’t see a lot of the time is the soft costs associated with manual work,” says Raise Robotics’ Oster. “Setting up fall protection, horizontal lifelines—these things are extremely cumbersome. They limit your work output, and resetting them constantly in between column spans eats up a lot of time.”

Being able to work from just behind a slab’s edge instead having a worker lying on their stomach isn’t just about increasing safety, but improving construction schedules and reducing risk as well, notes Oster.

Raise’s robot was initially designed to perform the layout, installation and inspection of fasteners and brackets for exterior wall panel installation on concrete slabs on mid-rise and high-rise building projects. But other use cases are already being developed since automating that particular set of tasks.

Raise is providing general layout and as-built inspections services with the robot as well. As facades are installed, the robot can document the conditions and compare them to project plans. “Most of the time this is for embed placement,” Oster says. “It allows you to get ahead of rework if you need to have a better idea on what the conditions are.”

Raise has already added a hammer drill attachment for the robot, allowing it to work alongside its contractor clients’ self-perform teams. Looking further into the future, Raise plans to add a concrete scanner attachment as well, so the robot can identify where rebar and post-tensioning are located so they can be marked on the slab.

“What we’re trying to accomplish with Raise is to create a multi-use robot, not for niche applications, but definitive applications,” Oster explains, saying there’s a sweet spot for jobsite robots. “Not being over-generalized, but not being too specific at the same time.”

pooling water observed by the robot’s camera

Gilbane’s client was concerned about windblown rain on certain floors of the project, so Nextera added functionality to note any pooling water observed by the robot’s camera.
Photo and image courtesy Nextera

Robots as a Gateway to AI Insights

When Gilbane Building Co. first began deploying robots from tech startup Nextera Robotics, it was always about expanding their digital platform. Nextera’s small, treaded robots can readily navigate jobsites, climb stairs and surmount or avoid obstacles autonomously to capture thousands of 360° photos, but that’s just the beginning of their usefulness.

“The biggest benefit is efficiency—we can get so many different outputs from one scan,” explains Ryan Spotts, Gilbane senior project executive. He is currently using Nextera’s robots and its Didge digital platform as the contractor builds the 17-story Children’s Hospital of Philadelphia Schuylkill Avenue Research Building. “We’re getting information on safety, quality control and progress monitoring, as well as keeping our client informed.” The Nextera robots are fully autonomous, deploying from their onsite charging stations to move through the multiple floors of the site, taking regular images room by room and hall by hall to create a digital repository of project imagery. With the building’s facade mostly completed and work shifting to interiors, this kind of site documentation is accelerating work during the current phase, Spotts says.

The robots typically operate on the second shift, so relatively few workers have to deal with them, Spotts explains. But even when workers see them rolling around, it quickly becomes a normal background activity they can generally ignore. “From the standpoint of the trades, this is not an audit tool used against them. It’s a proactive tool showing what has been installed.”

Gilbane has used manual reality capture tools in the past, but it was hard to justify the expense of sending a worker around to record site images on a regular basis, let alone once a day or even more frequently. Spotts notes Nextera has helped resolve disputes with subcontractors over what was installed when, and the project’s owner values the regular updates it generates. “People spend time on value-added problem solving, not data gathering, and smaller projects that don’t have the budget for robust documentation can get the same level of reports … I see this completely changing how we deliver large-scale work on campuses,” he says.

But the real benefit comes once Nextera applies its AI-based tools to all the captured imagery in its Didge platform. There are already tools to identify if workers are wearing PPE or if there are unsafe site conditions such as trip hazards or open holes, but Didge also offers progress tracking and BIM integration for checking if work is installed correctly. More importantly, it can offer AI-driven predictions on whether the construction pace is beginning to slip. And small corrections, thanks to these warnings, can add up, says Jacob Ryals, director of operations at Nextera. “All of these micro adjustments, day by day over a one-to-three-year project—the amount of time saved is huge. And on the safety side, rectifying hazardous conditions on a day by day basis and not finding out later in an accident: you’re reducing the time, money and personnel needed to rectify these issues over the course of a job.”

Once the robots start to show results, Spotts found that buy-in from stakeholders was relatively easy to acquire. And that’s the key element to making any deployment of new technology work, says Ryals. “[The user] doesn’t have to see everything, they can get super high-level stuff in Didge on a regular basis, but can dig in deeper if something is off. Five minutes after the robot captures something, you can get an alert.”

For Spotts, having the data available to make these kinds of decisions will change how he tackles complex projects with multiple subcontractors. “We find an issue with one subcontractor and can resolve it for another before it happens now that we have the data. We’re always using data through a problem-solving lens, so there’s been no pushback from the subs.”