Industrial Robot’s Joanne Pransky interviews Tessa Lau, CEO at Dusty Robotics about Dusty’s autonomous mobile robots that print layout plans onto the floors of job sites using information from building information models (BIM) as a guide.
Listen to this article
Tessa Lau is the Founder and CEO at Dusty Robotics, whose mission is to increase construction industry productivity by introducing robotic automation on the job site. Dusty’s FieldPrinter autonomous mobile robots print layout plans directly onto the floors of job sites using information from building information models (BIM) as a guide. FieldPrinter also generates status reports in real-time so project managers monitor progress and make adjustments if needed. In December 2019, Dusty closed a $5 million Seed round, bringing the total amount of funding that the company has received to $7.2 million.
Prior to co-founding Dusty in April 2018, Lau was CTO and Co-founder at Savioke, where she orchestrated the deployment of 751 delivery robots into hotels and high-rises. Previously, she was a research scientist at Willow Garage, where she developed simple interfaces for personal robots. She also spent 11 years at IBM Research working in business process automation and knowledge capture.
Lau was named 2017 Woman of Influence by The Silicon Valley Business Journal and one of the most creative business people by Fast Company in 2015. Over the years, Lau has served on program committees for various major HCI and AI conferences and on the board for the CRA-W – the committee for the status of women in computing research. Lau received her BA and BS from Cornell University in computer science and applied & engineering physics, and an MS and PhD degree in computer science from the University of Washington.
Joanne Pransky: Of all the robots you have worked with, what ‘misbehavior’ did you find to be the most interesting, unexpected or surprising?
Tessa Lau: Savioke Relay robots had so many misbehaviors that I think they had minds of their own. One of the most challenging tasks we had to do with the Relay was to program it to take elevators (Figure 2).
As humans, we don’t think about it, but there are all these little micro behaviors that we exhibit when we use an elevator and we’re using them alongside other people. That’s what we were teaching our robots to do; simple, expected human behavior like turning around and facing forward when it goes into the elevator, or exiting properly when the door opens at the floor it’s supposed to get off without dithering or waiting before deciding to get off. All of these little individual behaviors were a big challenge to get Relay to accomplish properly.
Joanne Pransky: Why did you decide to leave Savioke?
Tessa Lau: I was a co-founder of Savioke and had been there for five years. This is a really long time to stay with any company, especially at a startup in Silicon Valley. After five years, I realized that I wanted to start my own thing.
I’m more excited about the early parts of growing and building a company because there are just so many possibilities. It is a wide-open field. And there are no right answers, and you have to figure it out as you go. I really wanted to get back to that. So that’s how I started Dusty.
Joanne Pransky: How did Dusty get its name?
Tessa Lau: In the early days of Dusty we were trying to figure out what our first product would be, and we were doing a lot of market research. We had come up with a bunch of different ideas that we were testing with customers.When we got our first term sheet and offer from our first investor, we could not accept the term sheet until we had a company name, even though we hadn’t finalized our product idea.
We didn’t know we were going to be building this field printer. At the time, we thought we were going to be building a clean-up robot, a Roomba for construction that would actually come through and sweep up the dust and the dirt on the construction sites and leave it cleaner than before. On every single construction site, there’s always someone pushing a broom around, so it seemed like an obvious thing to do. The construction vacuum robot would come through overnight, like the cleaning fairy, and the workers would come in the next day and arrive at a clean site.
So we got that term sheet and the construction cleaning robot was the idea that we were going with at the time. We didn’t want to make the name too specific just in case we changed our minds later. Since the environment’s dusty and dirty, we came up with Dusty Robotics to indicate that our robots will be comfortable operating in the dusty environment of a construction site. Branding was really important to me from day one.
Joanne Pransky: What are some of the differences in your role as CTO of a startup compared to CEO and which do you prefer?
Tessa Lau: I have really enjoyed being a CEO. I wasn’t really expecting that, but it’s growing on me a lot, and I really love it.
As a CTO, my role was to oversee the engineering team and figure out what we are going to build now and in the future. It was a relatively narrow view. I didn’t get to see or have control over all the other aspects of the business.
As an engineer, I thought at the time that the CTO, the engineering part, was most important. It turns out, however, that the rest of the business is even more important. Now as a CEO, I get to learn about and master all of those other aspects of the business that have a huge impact on whether we’re going to be successful or not. I’m just having so much fun.
Joanne Pransky: Is Dusty the first company to develop an automated solution for the printing layouts in the construction industry or have there been other attempts by other companies?
Tessa Lau: We have competition. One of the reasons that I am confident we’re on the right track is because our customers have tried to build this robot before. DPR Construction was one of the top ten construction companies nationwide. Ten years ago, they tried to build a layout robot they called “laybot”.
DPR spent about six years and lots of money building this robot, but they never got to the point where they had it working in the field as well as we have. We’ve talked to the DPR people who worked on this and honestly, we’re trying to learn from their experience because they understand the market. They know why this is important and have already done the analyses to justify it. So as roboticists, we’re partnering with some of the most forward-thinking people in the construction industry that have all of these great ideas about how they want the industry to become more automated in the future.
Joanne Pransky: Conventional ink jet printers include a horizontal scan mechanism that transverses the paper. Is Dusty’s FieldPrinter (Figure 3) the scan mechanism, or does it position itself and then use on-board x/y scanning head to print a section of the drawing at a time?
Tessa Lau: That’s a great question. A typical printer is optimized for printing pages. Dusty optimizes for printing lines. What that means is that we’ve had to develop some new and specific solutions.
Robots are good at driving around on job sites, but they’re not so good at driving precisely. We’ve developed a hybrid solution that has our drive mechanism be the course control with an onboard printing stage that does fine control. The goal is to combine these two mechanisms in order to create a very precise, straight line.
Today’s state-of-the-art mobile robotics are accurate to within one to two inches and since we needed a 1/16 of an inch, we had to invent a new technology.
Joanne Pransky: Dusty is all about localization, positioning, and orientation. What can you tell us about how it does these three tasks?
Tessa Lau: One of the key challenges that we had to solve at the beginning of Dusty was trying to figure out how we could actually make the printer accurate enough to be usable on construction sites. How do we know where we are with enough detail so that we can print a line within 1/16 of an inch of what it needs to be?
Today’s state-of-the-art mobile robotics are accurate to within one to two inches and since we needed a 1/16 of an inch, we had to invent a new technology. The gist of it is that we’re making use of a piece of equipment that already exists on construction sites for doing very precise distance measurement.
Construction sites actually use a piece of technology called a total station. These are yellow tripods with a sort of turret on top, and it shoots a laser at a guy in the distance who’s holding a mirror and it measures that distance really precisely. We basically put that mirror on board our robot, and that device tells us exactly where we are within millimeters (Figure 3).
Joanne Pransky: One difficulty with marking up floors, etc., is that as one trade moves in they will cover up and rub out the marks that Dusty has created. How do you plan to get around this problem so that the follow-on trades can benefit?
Tessa Lau: Today, construction sites make lines by taking a piece of string and dipping it in chalk and then they stretch the string out between two points, and then they snap it. It leaves a line of chalk dust on the floor. But chalk is impermanent and it blows away. So they actually come through with a can of clear spray paint and spray over the chalk. That sticks it down and lasts for about a couple of weeks.
We can do the same thing with clear spray paint, but we’re actually using better technology than chalk. We’re using ink and with its high durability, ink sticks a lot better to the concrete. It doesn’t actually need that clear coat over the top of it to keep it permanent.
Joanne Pransky: Rather than printing lines on the floor, could Dusty instead project optical/laser lines on the scene – or have both capabilities? For example, could Dusty project the positions of light switches. etc., onto a partition wall?
Tessa Lau: Absolutely. Some of Dusty’s competitors are actually using projection-based layout solutions. One of them is a Canadian company called Mechasys. They have a device that instead of marking points and lines on the floor, they project them with the laser. Once it’s on the floor with the laser, people come through and build off of it.
Their markings only last as long as you have a laser up and running. If you turn off the laser or you lose power, then your markings go away. We obviously think our solution is better, but we do have some competitors that are already doing just that.
One of the things that we realized is a lot of information about what has to happen vertically, such as the height of a light switch, could actually be printed on the floor. That solves most of the problems that people have when they’re building. I think our solution actually gets us most of the way.
Joanne Pransky: What is Dusty’s business plan for the next ten years?
Tessa Lau: Ten years is a really long time! I can’t see that far out yet.
The plan for 2020 is we’re providing robots as a service (RaaS). What I learned about robots-as-a-service is that customers don’t want to pay for robots; they want to pay for value. If you can figure out and give your customers what’s valuable, that’s what they will pay for.
We’re actually operating layout as a service (LaaS). Our customers don’t pay for the robot. They pay us for the layout, because they need it. That’s how they build. Layout is a commodity people pay to create.
Joanne Pransky: What is your proudest moment of your whole career?
Tessa Lau: My proudest moment was when we did a pilot for one of our early customers in January 2020. It was going to be a free pilot as it was our first time engaging with this customer, and to be honest, our robot had been acting up lately. But it went perfectly, without a hitch. We had two days blocked out for us on the schedule. We could have finished it in a single day, but we actually slowed it down because they were bringing a camera crew the second day.
Our robot performed flawlessly. After we were done, our customer said, “You guys did great. You should send us an invoice.” That became invoice number one; our first paying customer. We had been building up to this. It was everything coming together – our engineering team, product, operations, our visionary customer – and making it all work.
Joanne Pransky: What is the biggest mistake or most valuable lesson that you’ve learned thus far in your career?
I spent the first six months talking to construction companies and trying to understand their problems and pain points before settling on our first idea and before we even bought our first piece of hardware.
Tessa Lau: There are so many. I’m fascinated by this process of creating startups and turning them into viable businesses. I learned a lot from the mistakes we made at Savioke.
When I left Savioke and started Dusty, the key thing that I learned was not to start building the product before you know who’s going to pay for it and why. As a result, I spent the first six months talking to construction companies and trying to understand their problems and pain points before settling on our first idea and before we even bought our first piece of hardware. That’s why we’re now relatively successful because we spent a lot of time upfront trying to figure out what is our product going to be and why it’s valuable.
Joanne Pransky: What do you think PhD and masters of engineering students should be doing while in school to prepare them for the commercial side of robotics?
Tessa Lau: I came out of academia. I have a PhD. A PhD teaches you some things that are really important, but not other things that are also really important.
The first important thing to learn in an academic environment is critical thinking. How do you evaluate ideas? If you have an idea of something to build, how do you test it and set up an environment that lets you evaluate it quickly? At Dusty, we’re running a set of experiments really fast, one after another. Will someone pay for this? If we tweak it this way, will they pay more for it?
The second important thing that I learned as an academic are communication skills, i.e., how to take something really complex and convert it into something that lay people can understand. What they don’t teach in a PhD program, and which matters even more than the engineering, is in terms of ultimate success. How is this technology going to be useful for someone in the real world? How do you critically evaluate a business idea? When you go work for a company – is their product something people will pay for? Thinking about that could drastically impact the trajectory of one’s career.
Joanne Pransky: What do you think you’ll be working on 20 years from now?
Tessa Lau: Twenty years from now, we’ll have robots doing all the work and hopefully, I’ll be financially retired. What I would love to be doing after Dusty is successful is to continue to create technology that impacts people and changes the world for the better by advising and mentoring the next generation of startups, particularly robotics startups.
Joanne Pransky has been an Associate Editor for Industrial Robot Journal since 1995. She was also one of the co- founders and the Director of Marketing of the world’s ﬁrst medical robotics journal, The International Journal of Medical Robotics and Computer Assisted Surgery. Pransky also served as the Senior Sales and Marketing Executive for Sankyo Robotics, a world-leading manufacturer of industrial robot systems. She has consulted for some of the industry’s top robotic and entertainment organizations, including Robotic Industries Association, Motoman, Stäubli, KUKA Robotics, ST Robotics, DreamWorks, Warner Bros., and for Summit Entertainment’s ﬁlm Ender’s Game, in which she brought never-seen-before medical robots to the big screen. She can be contacted at joannepransky[@]gmail.com.