Orr Fabian, the vice president of R&D at Skyline Robotics is dedicated to creating robots that help humans accomplish more and do things better, faster, and more reliably.
He previously served as co-founder & CEO of Inka Robotics, where the first autonomous tattooing robot controlled by computer vision was developed. The robot’s development focused on precision with the tattoo needle based on advanced computer vision: You only get one shot at doing a tattoo right and you have to factor in the diversity and complexity of human skin.
When Orr later began developing Ozmo at Skyline Robotics, that type of precision would come back into play. In order for Ozmo to be able to clean windows precisely and without error, it was essential that it be able to recognize the building’s façade and that computer vision and the robotic arm work together seamlessly.
Orr’s work has paid off as Ozmo robots are now cleaning skyscrapers in New York City and will soon be headed across the pond to London.
We invite you to this TechBullion interview with Orr Fabian about his robotic journey to Skyline Robotics and how he continues to push the boundaries of robotics to benefit the world.
As the vice president of R&D at Skyline Robotics, please enlighten us on what is inspiring this dedication to push the boundaries of robotics to benefit the world?
It was my childhood dream to change the world using technology, solve complex problems, enhance efficiency dramatically, and save lives if I can. Ozmo, the first window cleaning robot, allows me to do just that in an industry that needs to be automated. The window-cleaning industry faces many challenges. First, it’s an antiquated industry. Not much has changed in window cleaning in the last 100 years despite buildings getting bigger and higher. That brings us to the second challenge: Safety. Window cleaning is a dangerous business. Workers are suspended on a platform outside of a building. That may not have been as big of an issue 100 years ago, but again, the sheer size and volume of buildings today have made it incredibly dangerous. Finally, the long-term outlook shows a severe shortage of window cleaners. Three quarters of window cleaners in the USA today are over 40 years old and only 9% are between the age of 20 and 30. Meanwhile, new buildings continue to be constructed. These issues inspired the creation of Ozmo. With Ozmo, we can leverage robotics to make window cleaning faster, safer and more efficient.
You previously served as co-founder & CEO of Inka Robotics, could you give us a recap of how this journey all started with an overview of this technology?
The idea actually came from a student in Auburn University named Luke Gehron. He took a 3D model of a drawing and used a 3D printed machine to draw that model with a marker on his arm. This immediately got me thinking, “why not tattoos?” I started Inka Robotics in 2014 with a mission to create a tattooing robot—which at the time I began—was naïve in terms of what I thought I needed to accomplish, both business wise and technologically. But the insane opportunity was there; in the USA there are more people with tattoos than people who own iPhones, not to mention this industry hasn’t changed in over 130 years.
So first thing I had to research how to tattoo, tattoo statistics, clients, artists, shop owners. I went to almost every tattoo shop in Israel (ended up partnering with the biggest one). I understood that I needed to create something of true value to the customers and the end users. Something that doesn’t leave a gap between the client’s desire and the actual product. Something that assists the artists rather than replace them.
Apart from building a team of experts, I had to partner with industry leaders (such as with Yaskawa, a robotic arms manufacturer) to create a robotic arm with a tattooing machine at the end effector, and a depth camera for computer vision. This tattooing robot was much safer, less painful, more hygienic, far faster, more accurate, and gave a real 3D preview of the tattoo on the skin.
The idea was really innovative—we used depth cameras to detect & track user movements, and to continuously update the 3D path of the robot. The human skin is a deformable object—it keeps moving, a person breathes, muscles move, you’ve got to have intense real-time control over the tattooing needle. Remember, this is a tattoo so there is no room for mistakes!
In this process I also realized I’m creating a digital tattoo ecosystem, an ecommerce system where everyone can upload their tattoos including famous artists. So if someone can’t reach his favorite artist, he can just use the robot in the nearby shop. This is an example of how robotics is always more than the original application it was created for.
Our readers would love a sketch of the world of robotics, what is the global market size of this industry and what are the next big trends to lookout for?
The global market size of the robotics industry was valued at around $45B in 2020 and is expected to reach over $147B by 2025, growing at a CAGR of approximately 20%.
The biggest trends in robotics will probably be:
1) The adoption of collaborative robots, designed to work safely alongside humans, is on the rise. These robots offer increased flexibility and can assist in various tasks, enhancing productivity and efficiency across numerous industries, including real estate (like Ozmo).
2) Integration of AI and ML technologies into robotics is becoming more prevalent. Advanced algorithms enable robots to learn and adapt to their environments, enhancing their capabilities in perception, decision-making, and autonomous operation.
3) The development and deployment of autonomous vehicles, including self-driving cars, trucks, and drones, continue to gain momentum. These technologies have the potential to revolutionize transportation, logistics, and delivery services.
You have been instrumental in the creation of both, the world’s first window cleaning robot, and the world’s first tattooing robot, what do these 2 robotic projects have in common?
In a word: Precision. A key area of focus in the tattooing robot’s development centered around precision with the tattoo needle and the computer vision: You only get one shot at doing a tattoo right and you have to factor in the diversity and complexity of human skin. When I later began developing Ozmo at Skyline Robotics, that type of precision would come back into play. It was imperative for Ozmo to be able to respond to a building’s façade and that the computer vision and robotic arm work in tandem for precision window cleaning with no mistakes.
What is precision computing, how does it work and what role is it playing in your robotic projects?
Precision computing refers to the use of high-precision calculations and algorithms to enhance the accuracy and reliability of computational systems. It involves performing calculations with increased precision beyond what is typically used in standard computing operations.
Precision computing can handle extremely large or small numbers, maintain precision over extended calculations, and reduce the cumulative effect of numerical errors.
In our project, we use it for sensor fusion (combining data from several sensors, such as the lidar, IMU, distance sensors, and more to increase certainty during measurement), motion planning (having the robot perform precise and smooth movements when cleaning), localization (when the robot needs to understand its position relative to the facade), and computer vision (detection the window sills more accurately for example).
This robotics journey led to developing Ozmo at Skyline Robotics, the next big thing in window cleaning. Could you give us a recap on Ozmo and your success stories so far on this project?
The first Ozmo robots have been fittingly deployed in New York City, the mecca of skyscrapers. We have partnered with Durst Ventures, an affiliate of The Durst Organization, that owns some of New York’s most notable skyscrapers including One Bryant Park, One World Trade Center and 151 West 42nd Street. We have also been awarded patents in Japan and Singapore as we begin to expand our business globally. Additionally, we have received several awards including an RBR50 Robotics Innovation Award.
For those who might think using Ozmo will be challenging, could you give us a walkthrough of the Ozmo system?
Ozmo is a state-of-the-art robotic system for skyscraper window cleaning. Ozmo uses the existing building’s infrastructure in order to operate. On every skyscraper you have what you call a BMU (building maintenance unit), a giant crane that carries a large basket. Usually, window cleaners stand in the basket working in the beating sun, a very repetitive job, not to mention dangerous. Ozmo is there to send them to the roof, to a nice UI, and to be in the basket instead of them. Ozmo has an articulated 6 DOF robotic arm that mimics the window cleaner motion, it uses purified water—reverse osmosis water (that’s where the name Ozmo comes from!)—so the water won’t have any minerals inside of them (to prevent water spots on the window). Ozmo has all kinds of sensors on the basket and on the roof to monitor its own work. The LiDAR, for example, creates a point cloud—it’s Ozmo’s way of localizing (or understanding its own location in relation to the façade itself) and detecting window sills & the window plane. It also allows it to work in complete darkness if we want. Ozmo has a force sensor (to measure how much pressure is applied) & distance sensors (some laser, some ultrasonic—detecting distance from the window) to detect the window plane and orientation, and to detect the sills as a redundancy. Ozmo also has an IMU to detect the orientation of the basket. It also has load cells/tension sensors for the water hose and the electric hose, and a water system to control the pressure & flow of the water (to avoid dangerously high pressure and to apply just the right flow for perfect cleaning).
It seems your hardwork has paid off as Ozmo robots are now cleaning skyscrapers in New York City. What is next for Ozmo, do you have any plans of global expansion?
Yes, Skyline Robotics is in the midst of a global expansion. Again, this year we have already received patents for our Ozmo window-cleaning systems in Singapore and Japan. Go-to-market strategies are now being implemented for major cities across the world, partnering with regional, market-leading window-cleaning companies.
AI and Robotics projects are the next big ventures with investors flocking to cut a share of viable projects like yours. Any available opportunities for investors and partnerships you would like to share with us?
We’ve had quite a bit of interest from investors and potential partners based on the proven success of Ozmo. We are always open to discussing opportunities that will further our company’s mission to bring Ozmo to skyscrapers globally. Interested parties can contact us here.
Do you have more tips to share with our readers today?
Replacing “regular muscles” with “mechanical muscles” always makes perfect sense when it comes to robotics. The fact that you can clean 3 times faster, and at night, means our customers can clean more buildings during the cleaning season. Less people per system means cleaning more buildings at the same time. The more efficient Ozmo becomes, the bigger the incentive will be to use it, including in markets where the labor costs are very low.
The current window cleaning labor force will transition to better jobs, including operating the robots, inspecting the work in a GUI, far from harm’s way, and far more interesting.
In the future Ozmo could be used for a bunch of other applications, including façade health, maintenance in other areas, and in other industries.