While Amazon won’t participate at CES 2022 in person in Las Vegas to prioritize the health and safety of our employees and our partners, we’re excited to highlight partners like Labrador Systems who are introducing new and innovative products for customers.
The company’s Labrador Retriever is a new type of personal robot that physically helps people move large loads around the home (such as the laundry basket, groceries, or dishes for the table) and keeps essential items within the reach is an exciting manifestation of this vision of ambient intelligence.
The Retriever is developed by Labrador Systems, an early-stage robotics company pioneering a new generation of assistive robots. Amazon’s Alexa Fund is an investor in Labrador Systems, which has been awarded a National Science Foundation grant to fund its work into developing new technology for enabling assistive robots for personal care and COVID-19 protection.
Mike Dooley is the CEO of Labrador Systems. Dooley first experienced the problem that Labrador would go on to tackle when his parents visited him for an extended vacation.
“My mom had just started using a walker,” Dooley says. “She was finding it difficult to perform basic tasks like carrying the laundry or bringing my Dad a cup of coffee. Seeing this scenario repeat for a few weeks was hard, as she was in that zone of wanting to be very active, but these gradual changes were progressively having a bigger impact on her daily activities.”
Several years later in 2017, Dooley co-founded Labrador Systems. During that time, the technologies available to robotics had started to make significant leaps, including the advances of 3D vision systems used in augmented reality (which can operate on low-cost cameras, and help a robot understand its location), the improvement in depth-sensing cameras (which are critical to moving a large robot around a complex environment like the home), and the development of the Robotics Operating System, called “ROS” (which enables developers to more rapidly build a variety of applications on a common architecture).
In the course of his career, Dooley has been involved in some of the significant announcements in the field of robotics at large. As part of a skunkworks project at LEGO in the late 1990s, Dooley helped bring LEGO Mindstorms to market. Later at Evolution Robotics, Dooley and Labrador co-founder and CTO Nikolai Romanov were part of the team that launched the Mint series of smart robotic floor cleaners in 2010. Evolution Robotics was acquired by iRobot, and Mint evolved into what is now iRobot’s Braava product line.
In the conversation below, Dooley spoke about some of the technical challenges the team had to overcome while developing the Labrador Retriever, why he decided to incorporate Alexa into the product, and how the Labrador Retriever is empowering individuals to lead more independent lives.
“From a scientific perspective, the most significant challenge that we had to overcome was to build a robust navigation system for the Labrador at an affordable cost.
Most commercial robots that self-navigate with indoor environments use LIDAR (Light Detection and Ranging) as their primary sensor for their localization system. As you might know, LIDAR systems use lasers to scan and measure distances of the surrounding walls and objects that are around the robot.
LIDAR has been used for in years in commercial SLAM systems (Simultaneous Localization and Mapping) to map and navigate spaces. It is a proven and reliable solution for warehouse and other commercial applications. However, high-quality industrial grade LIDAR is simply cost-prohibitive for what we are trying to accomplish.
On the consumer side, several of the popular smart robot vacuums available today use visual SLAM to navigate spaces. Some vacuums a use low-cost version of LIDAR. However, those sensors frequently result in a trade off when it comes to quality and resolution for position accuracy. The vision systems on these robot vacuums use one or more low-cost cameras, like ones you can find in a smart device, to collect visual data points and create a map of their surroundings. By relying on a small and readily available piece of hardware, they can be incredibly cost-effective.
When we looked at our requirements, we fell in the middle. We needed something similar to the level of performance of a commercial navigation system. We also needed to achieve the low cost of the components used in consumer electronic products. At the time, there were no solutions in the market that addressed our needs.
We saw really promising technology for vision-based localization coming out of augmented reality (AR), which could run on consumer devices, like in smart phones and headsets. However, those applications weren’t designed for scenarios that involved a robot in someone’s home. If you are playing a game on an AR headset, the system can ask you to turn on more lights on in your room if the visibility is too low. We don’t have that option. Our goal is to train the robot on the environment once, and make it adaptable to operate under all sorts of changes in conditions.
To give just one example, a common story we hear in our research is someone who goes to bed, and wakes up needing something. If they are using a cane or a walker, they often would prefer not to get up in the dark. In these cases, our mission is for the robot is to bring what they need to their location, which can often be engulfed in complete darkness. As a result, we’ve spent tons of time training and testing the robots in different lighting conditions, so that a robot that was trained in bright daylight still works when it gets the call in the middle of the night.
There were several other challenges that we had to overcome, and it would take me hours to go through every one of them. What I will say is that the people that have been testing the robot in their homes have been a critical part of the motivation that keeps us going. It’s really heartening to see how fast they adopt the robot, and the difference it makes in their daily lives.
I’d also like to add that incorporating Alexa has been critical to fulfilling our mission of empowering individuals to do more of what they want to do. People can now perform everyday tasks by using their voice. For individuals with pain or other health issues impacting their daily activities, that can be an incredibly liberating experience.”
“The Amazon Alexa Fund is an investor in Labrador Systems – but that’s not why we decided to go with Alexa. We chose Alexa as the intelligent voice assistant for the Retriever, because it is the gold standard for intelligent voice assistants. You have so many developers and brands building a wide variety of skills and customer experiences on the Alexa platform, and we are now able to access this entire ecosystem through a single point of connection.
The ease of development was another important factor that influenced our decision to go with Alexa. Our team excels at robotics. However, we are not experts in natural language understanding. Alexa’s conversational AI platform allows us to draw on all the advances in NLU, without having any expertise in the area. The development process was incredibly easy: we went from concept to a fully functional implementation in two months with just one engineer working on the integration. What’s more, with Alexa, things like security and privacy are already taken care of, which is especially important for products developed to function in people’s homes.
I would encourage businesses looking to incorporate voice and enable new customer experiences to build on the Alexa platform. Building something from scratch is not easy, and you shouldn’t underestimate the work that people have already done. Don’t reinvent the wheel if you don’t have to.”
“When it comes to robotics, I approach the future with a sense of guarded optimism. With all the coverage in the media, it’s incredibly easy to buy into the hype. However, it’s important to recognize just how far we are from having R2D2s and C3POs.
I think it’s far more productive to think of the future in terms of how we can develop and apply technology to empower people to do more, whatever that means for them. That’s been my guiding philosophy throughout my career. The LEGO Mindstorm consists of a relatively basic set of sensors, motors and interfaces – but it became one of LEGO’s bestselling products of all time. It introduced so many people to robotics, and did something really important for society. You can say the same for the development of the breakthrough technologies at Evolution Robotics: they advanced the state of the art in a profoundly meaningful way.
That’s why I’m so excited about launching the Labrador Retriever with Alexa at CES 2022. This is a new chapter in robotics. I think about one of the participants from our home trials. She is fighting multiple health issues. She told us about the Labrador Retriever and Alexa are helping her perform everyday tasks, so that she had the energy do the more important things in life. The word that kept coming up was “independence” when she described her experience. To see that kind of impact is what gets me excited about the future of robotics.”