How a machine learning scientist from NASA is accelerating digital transformation for Architectural Engineering and Construction (AEC) professionals
Creating accurate floorplans used to be a painstaking, manual process with a steep learning curve; for most applications, the toolkit comprised a tape measure, a clipboard, and ample tenacity. Today, AI-powered mobile solutions like the one from California-based OpalAi allow Architectural Engineering and Construction (AEC) professionals to capture a detailed 3D floor plan in minutes with minimal training, drastically reducing the time spent measuring and creating as-built models. OpalAi’s next-generation artificial intelligence processes your 3D scan into an editable, professional-grade CAD model for downstream use in various architecture and design software, including Revit, SketchUp, AutoCAD, and Rhino.
During the Covid-19 pandemic most of the people around the world spent their time at home due to the pandemic. As a result of that many people understood the value of home and the environment more than before. Many people, especially millennials, rethought about the housing situation and decided to move far from the metropolitans and live in larger houses rather than renting a small apartment in a metropolitan area.
Dr. Ryan Alimo, a millennial and a machine learning scientist at NASA’s Jet Propulsion Laboratory (JPL) California Institute of Technology (Caltech) and the founder of OpalAi, while doing a home remodeling project noticed the huge cost for such a simple task, creating an as-built drawing or floorplan (“floorplan” is the bidimensional spatial arrangement used by designers to determine internal building layouts, also sometimes referred to as blueprints, spatial layouts, or internal spatial arrangements) of an existing building. After discussing this issue with his sister, Audrey Alimo, a California Licensed Architect (AIA), he found out that there is a significant market need for autonomous low-cost floorplan reconstruction for existing residential and commercial buildings. Initial phases of an architect’s costs for tenant improvement projects, which includes floorplan reconstruction, accounts for 10%-25% of the architectural costs in the U.S. residential and commercial market (Ebhomien, 2019). More than 90% of the houses in North America are missing accurate floorplans; in 2018, an estimated 13 million renovation projects were completed which cost homeowners over $1,000, of which roughly 20% hired a design professional to generate the floorplan (Bustamante, 2020).
With the request from his sister Audrey, principle at Alimo Architects, Ryan (founder of OpalAi and machine learning scientist at NASA’s JPL) started to prototype the idea and generate the autonomous floorplan of the buildings. They found that in order to guarantee the level of accuracy required by the architects (usually 99 % of human measurements or 1 inch error) it is required to use a depth ranging sensor such as Laser or LiDAR. Observing the technology trend, they predicted that consumer depth sensors such as LiDAR scanner would be widely available in the coming years, which a similar technological revolution happened to GPS sensor in 2007. OpalAi is sponsored by the National Science Foundation (NSF) Small Business Innovation Research (SBIR) to create an autonomous floorplan generation leveraging multi-purpose commercially-off-the-shelf (COTS) hardware (e.g., iPhones 12-14 Pro, iPads Pro), used by the users themselves. OpalAi’s product, ScanTo3D application, enables users to inexpensively generate 3D views and floorplans without specialized hardware and personnel training. This is unlike many competing products, available primarily to big firms, which rely on expensive hardware with many manual steps to produce accurate results. The Economic Value to Customers (EVC) is 10%-25% of the entire project fee in architectural works. This mobile based solution substantially reduces the soft cost in home remodeling, allowing small and midsize architectural firms to minimize their variable costs and thus be competitive with much larger firms, leveling the competitive playing field.
Almost 70% of the architect’s costs are from transportation, referral, and design documents. ScanTo3D saves at least $1000 for every scan resulting in a 10 times return for AEC professionals with every purchased of floorplan using OpalAi ScanTo3D.
According to the National Association of Realtors (NAR) research survey, 55% of home buyers believe that floorplans are very useful features on real estate websites (Cone, 2018). It is speculated adding a floorplan to real estate listing on a website can increase the click-through rate by 52% (Fisher, 2020). There is a significant market need for floorplan generation as 5.34 million houses were sold in 2019 (NAR, 2020).
ScanTo3D product reduces the costs involved, saves time spent on traveling to job sites, reduces human error, and rapidly creates high-fidelity digital copies of physical places. This will have real and material impact for Americans and will raise the relevant sectors’ interests.
Market need
There is a significant market need for autonomous low-cost floorplan reconstruction for existing residential and commercial buildings. Currently, most of the work to create floorplans is performed manually; it is time-consuming and very expensive. Based on interviewing close to 100 potential customers in Southern California, OpalAi estimated that for an average house with a size of 2500 sq. ft., it takes on average 40 hours by two employees to take measurements and to create the CAD model and costs ~$1250. For larger commercial and industrial buildings, this process can take weeks. For example, an average 100,000 sq. ft. space would take almost a month to create a floor plan and may cost up to $10,000, depending on the complexity of the space. The costs to hire a floorplan designer range between $50 and go as high as $130 per hour for a draftsperson to draw up blueprints or a space, while the costs to hire an architect range from $100 to $350 per hour (Forbes, 2022). For commercial building plans, draft fees are usually between 2% to 15% of the total construction cost (Home Guide, 2021).
Yossi Kahlon, CEO of Skand.io that provides spatial data management and visualization solutions and facilitate other companies to perform energy efficiency renovations projects mentioned that: Depending on the size of the space usually 1 or 2 inspectors will be on site to take measurements. It would take a couple of days to take measurements manually. Then it takes a day or two to input the data into the software and do some data validation. After that, a design team takes over and it takes them about a week to create CAD files. Estimator team takes 2-3 days to get cost estimates for the project. Thus, we can use up to 3 different teams for the projects. The quickest turnaround we have seen for some of the simpler projects would be about a week for the whole process, but since their clients mostly work with larger commercial and industrial spaces it is not uncommon that the whole process will last almost a month.
With the buildings sector accounting for 40% of global emissions, WorldGBC Net Zero Carbon Buildings Initiative calls for reducing emissions from buildings globally by 50% by 2030 and reaching net zero life-cycle emissions by 2050 (WGBC, 2021, AIA San Diego Design Week with a focus on Climate Action). Floorplans are an essentialstep for performing energy efficiency renovations. However, according to a CBECS U.S. EIA survey from September 2021, the median age of commercial building in the U.S. is 36 years, and buildings older than 20 years are most likely missing floor plans,aka as built drawings (CBECS U.S. EIA, 2021).
Thus, the larger the space, the more apparent the benefits for the companies to choose autonomous solutions to create floorplans. Not only that it becomes 30X times faster compared to traditional manual solutions, but it also directly translates into personnel cost savings, when companies can send fewer personnel and lower skilled personnel on sites. Construction and engineering firms are likely to improve their efficiency by using standardized products and software-based design automation. Time and cost savings also help the companies to close the deals faster and win more projects, as many construction companies use floorplans in an internal project cost estimation tool before they send an invoice to their respective client. However, as our customer discovery showed, construction remains a relatively low-tech industry, with most companies missing internal capabilities to develop software solutions for the floorplan creation process. That is why the B2B construction segment that works with large industrial and commercial spaces is where we expect to bring the most value to the clients. As outlined further in this commercialization plan, it is going to be our main long-term target segment.
What Do AIA San Diego members think about OpalAi?
OpalAi is endorsed by several AIA San Diego senior members including Steven Shinn, FAIA, the former president of the San Diego Chapter of the American Institute of Architecture (AIA), and Benjamin Regnier Design Technology Lead at Gensler.
OpalAi’s technology seemed to be a great additional tool to assist architects in reality capture and project start up. Benjamin Regnier has tested the ScanTo3D app with Ryan Alimo and found it intuitive, fast, and reliable.
Benjamin thinks that solutions like OpalAi products can benefit AEC professionals in several ways:
1) Turnaround time: Normally, we create floorplans through point-to-point measurement, or by manually analyzing point cloud data to create BIM and Revit models. Processing 3D LiDAR data with AI algorithms would be a much faster way to get CAD/ BIM models of spaces.
2) Communication: OpalAi’s solution can significantly improve communications with the contractors through a better understanding of current site conditions. It can be a useful tool to monitor the progress of the projects, receive real-time feedback from job sites, and comparing designs with conditions in the field.
3) Data privacy: Data privacy is a crucial factor for some of the clients. Users prefer most calculations to happen on the device and limited data to be sent to the cloud servers to protect the privacy of our customers. The fact that OpalAi’s software can generate a 3D point cloud model of the space right on the user’s device is very important factor to many architectural firms.
OpalAi is a technology startup in the southern California and AIA San Diego recommends their solutions to be tested and support this local Ai startup that is founded by UC San Diego Alumni.
1.1. How does it work?
The OpalAi ScanTo3D app measures and digitizes the built environment by converting point cloud data from a LiDAR-equipped mobile device’s camera sensor into a scale-accurate 3D replica of the area scanned.
All you need to do is download the ScanTo3D app with your iPad Pro or iPhone/ Pro 12 or newer versions and start scanning. No need to move any furniture or get the whole building in one scan, either – if the space is too large, you can capture it in multiple scans using the “extend scan” feature. You can even scan the exterior walls and site improvements for complete as-built documentation. When finished scanning, select your desired output format and upload the scan to our secure cloud server for processing – within a few hours you will receive a professional floorplan with measurements.
1.2. Who is it for?
Short answer – anybody that’s ready to ditch the tape measure and clipboard.
Until recently, 3D scanning was out of reach for most AEC applications involving existing buildings, especially residential remodelers. With OpalAi, no job or estimate is too small for a professional as-built floorplan. Whether it’s a simple bathroom remodel or a full-blown passive house retrofit, OpalAi has you covered. Save time and never miss another measurement, all while impressing your clients with cutting-edge technology!
OpalAi’s mobile scanning solution is suitable for projects of any size, including residential, commercial, and industrial spaces, both interior and exterior. Using the most advanced AI and machine learning tools will make creating floor plans for your properties a breeze.