Automating the drafting process
The potential of drones was evident fairly early. They were small and could fly, and being relatively affordable, they allowed businesses of all sizes to survey their properties, even the parts that were once inaccessible, in a short amount of time. But like with much innovation, the initial upside can create its own problems. Making information gathering easier is an important development, but without supporting technology, the datasets more than likely will sit on a computer or in the cloud. As David Morczinek says, “It’s pretty and accurate but what’s the impact of that data? What do you do with it?”
That was the genesis for AirWorks, the company Morczinek started with Adam Kersnowski in 2018. The startup takes data that surveyors and engineers collect through remote means, such as drones, sensors or lidar scanners, and through AI-powered software, converts it into CAD drawings. The big value-add is speed. “Surveyors and civil engineers are able to finish their work faster because of the automation, and having that data available earlier in the process allows stakeholders to make better decisions,” Morczinek says. That means a boost in overall efficiency. Projects start out stronger and require less rework as they progress through the design and build lifecycle. The end result is a reduction in costs, a straighter line to completion, and a building, bridge or roadway that closely lines up to what was originally hoped for.
Surveyors and civil engineers are able to finish their work faster because of the automation, and having that data available earlier in the process allows stakeholders to make better decisions.
Morczinek, an aerospace engineer by training, was a student at the MIT Sloan School of Management when, in the summer of 2017, a classmate introduced him to Kersnowski, who had a background in construction and unmanned aerial vehicles. When they met, they both wondered why repetitive surveying was needed and if there was a way of making the job easier and faster.
As it stood, a surveyor would go to a location, spend two hours and gather 500-1,000 points of interest, and then interpolate between any two. Information would be pieced together to create a more complete model of the surface, Kersnowski says. Surveyors could do a good job, but the shortcoming was that there was no technology to ensure that the data accurately represented reality.
Drones came in and helped to fill the gap by gathering both quantity and detail, but there still wasn’t any technology to do anything with all the intake, which led to backlogs of unprocessed data, he says.
This was the motivation for AirWorks. Morczinek and Kersnowski saw the opportunity to use automation to help surveyors and civil engineers leverage remote sensing and complete these jobs better and faster. As they both say, their technology can take what has already been gathered by drones as hundreds of millions of data points, and define and classify them into concrete feature classes, such as sewer covers, utilities or other assets. For aerial data, “there is no in between,” Kersnowski says. “There’s a better robustness to the data.” The rendering can then be stored and downloaded whenever desired by the client.
Costs and shutdowns often associated with surveying roadway and highway projects would be reduced, but, more than that, the data helps civil engineers understand the properties they’re working on more completely, Morczinek adds. The design process becomes smoother; more informed decisions can be made from the outset. “Better data early on has a significant value. It can cut cost of design changes dramatically,” Morczinek says. And in the end, when what was built is checked against what was intended, known as the as-built, both match up with little variation.
The company’s current focus is producing two-dimensional drawings of 2D and 3D data. The scalability is already good, but, to improve that, the next goal is to offer an interactive platform that allows users to “easily draft over aerial datasets,” Morczinek says. The target launch for that product is later in the summer of 2021.
As the world moves to renewable energies and shifts how people travel, roads and highways will need to be redesigned to fit the changing landscape and space needs.
Ultimately, Morczinek says that AirWorks wants to have software that can autonomously interpret data at the scale of cities or entire states. It’s the natural progression of what clients will need and projects would require. In just one example, Kersnowski says that as the world moves to renewable energies and shifts how people travel, roads and highways will need to be redesigned to fit the changing landscape and space needs. For such a complex undertaking to be done with minimal disruptions, up-to-date imagery and data will be key, and it’s what the AirWorks platform is ultimately suited for, he says.
Like with any kind of automation, AirWorks’s technology brings up the worry that machine learning will eventually replace the efforts of a person. Morczinek and Kersnowski understand the concern, but they say that their computational models first and foremost give more accuracy in less time and for less money. They want to make surveying easier and provide better data across the value chain. They have no intent to replace anyone.
It also wouldn’t be possible, they say. Project aspects like quality control and oversight can’t be automated. And creativity cannot be automated. The AirWorks software could produce a CAD drawing for a construction project that would be functional and usable, and it could still be something that no one would want to buy because it lacks sensibility, warmth and even style, all the things that only come from people being involved. The technology gives better data for better decisions, but those decisions still are made based on experience and expertise. “People are going to want special designs,” Morczinek says. “That requires creativity, and creativity cannot be done without human beings.”