To experience the state of the art in autonomous mobile manipulation, you’ll want to find some well-funded academic lab to visit. Or maybe check out Google, or Amazon, or Toyota Research, or drop in on the RoboCup@Home competition. Really, the only other place you’re likely to find an autonomous mobile manipulator is in a relatively structured environment in a factory or warehouse, and even that is pretty rare. Mobile manipulation is superhard, especially when you try to do it in a less structured environment, which may be full of all sorts of horribly unpredictable things (like humans).
Diligent Robotics, a startup founded by Andrea Thomaz and Vivian Chu, is undaunted by the challenge of autonomous mobile manipulation in semi-structured environments. Over the past year, they’ve been testing a one-armed mobile robot named Poli in several hospitals in Austin, Texas, where it’s learning how to help nurses with simple fetching tasks. And when we say simple tasks, we mean navigating through hallways full of busy people, picking up supplies, not dropping those supplies, and then navigating back again without running anyone over. In other words, from the robot’s perspective, not simple tasks at all.
Diligent has kept a relatively low profile in the robotics startup world, but just a few weeks ago, they announced a US $2.1 million seed round. Here’s a video from December 2016 showing a Poli prototype undergoing tech feasibility testing last year at Seton Medical Center at the University of Texas in Austin:
If you didn’t catch that, Thomaz pointed out that expensively trained and highly skilled nurses spend around 30 percent of their time on tasks that could be called “hunting and gathering”: searching for supplies and then bringing those supplies to where they’re needed. This isn’t what the nurses want to be doing; it’s not what the patients want the nurses to be doing; and it’s not what the hospital wants the nurses to be doing. It’s just one of those things that has to be done, and it sucks that it has to be done by humans.
Diligent wants robots to take over these tedious, time-consuming logistical tasks, letting nurses spend more time doing what they do best—interacting with patients. The robots won’t be taking over anyone’s jobs, and they won’t be getting in between patients and medical professionals, they’ll just be doing the dull and repetitive tasks so that humans won’t have to.
This isn’t a unique idea—robots are particularly good at taking over boring jobs so that humans can do what humans do best. And there are plenty of examples of robots doing that, even in hospitals. What’s different about Diligent is that their robots aren’t designed to take over any part of hospital logistics completely; instead, they’ll collaboratively work alongside humans in environments designed for humans.
Photo: Diligent Robotics
“The core focus of our company is on getting mobile manipulation to be successful in semi-structured environments in which the robot has to perform its actions alongside people,” Diligent cofounder and CEO Andrea Thomaz told us. “We think that human-robot interaction is going to be key to success, and we like to say that we focus on the suite of social-intelligence software that will be needed for these dynamic human environments.”
Potentially, this approach could make Diligent much more flexible than more traditional robotic logistic systems, and initial feasibility testing has shown that Diligent’s software allowed their robot to come into a new hospital and adapt its skill set to different tasks, including helping stock supply shelves outside patient rooms and manipulating objects found in the supply closet of a typical nursing unit.
For more, we spoke with Thomaz via email.
IEEE Spectrum: Practical, autonomous mobile manipulation in semistructured environments is a very difficult problem. What made you decide to start a company in this space?
Andrea Thomaz: We think that these are going to be the robots that truly realize the most value to customers. When robots can autonomously move around a space and perform manipulation tasks to cause changes in the environment, this dramatically changes the scope of problems that we can start to think about applying service robots.
Why is human-robot interaction so important for robots to be able to work successfully in environments like hospitals?
Humans are one of the main reasons semi-structured environments are hard. People create a lot of dynamics that a robot operating in the space has to account for. Our technical perspective is that modeling and forming expectations about human interactions in an environment is a fundamental skill needed to work successfully in semi-structured human environments like hospitals. Moreover, we believe that an HRI approach, of getting prototypes out in front of customers for rapid prototyping and iteration is going to be key to creating a solution that really works.
What kind of reactions did you get from the nurses working with the robot? How has that shaped your approach with your next round of pilot customers?
We have done feasibility tests at three different hospitals in Austin, and each time we get a very similar reaction from nursing staff. An initial skepticism/worry: “What is that for? Is it going to take my job?” Then when we explain that we are building the robot to play a supporting role, to do some logistical tasks that their time away from patients, the reaction quickly turns to “Wow, that would be fantastic, when are we getting them?” This is very important to the vision of our company, we are building service robots to play a useful role alongside teams of people.
Photo: Diligent Robotics
With seed funding in place, Diligent is working on an updated prototype of Poli, pictured here with cofounders Thomaz and Chu. That’s a Segway base, and once completed, it’ll include a Kinova arm as well, along with “a few other mods,” we’re told. The skins for the first Poli prototype, as well as the new version, are designed by Carla Diana, who also came up with the look for Simon.
In 2018, Diligent will be hiring an engineering team and deploying robots with their first group of pilot hospital customers. Their goal is to get a better understanding of what specific tasks are the most valuable, and to see how much of a difference service robots can make over time.
We’re looking forward to seeing the results of some of these pilots, which should provide a more specific examples of the kinds of autonomous manipulation skills that Poli is developing. And if Diligent can get Poli to work in hospitals, my guess is that they won’t have any trouble expanding to other environments as well.
Evan Ackerman is a senior editor at IEEE Spectrum. Since 2007, he has written over 6,000 articles on robotics and technology. He has a degree in Martian geology and is excellent at playing bagpipes.