On April 15th, Dr. Tra Vu, COO of OhmniLabs and Kambria, met with Pranali Desai, Co-founder of Robomechtrix, for an AMA hosted by Women in Robotics, a professional network within Silicon Valley Robotics. Below is a recap of the Q&A between Pranali, Tra and Women in Robotics members discussing COVID-19’s impact on robotics along with other facets of robotics development.
In her current role, Tra is scaling OhmniLabs’ robotics and AI products and services, and further expanding the reach of the Kambria open innovation protocols platform. Before joining these two companies in Silicon Valley, she was an adjunct professor at New York University Tandon School of Engineering, and a Technical Director at Greenman-Pedersen, Inc. (GPI).
If you could not catch the AMA live, please post your questions for Tra on Twitter. The team will get back to you quickly with a response.
How have you been? How is working from home going?
I am doing well, all things considered…safe and healthy. Our team is still working hard — building telepresence robots is considered essential work. We’re also making face shields for donation to local hospitals thanks to Helpful Engineering’s guidance. So, yes, most of my team, myself included, is working at our robotics facilities.
Wow, that is fantastic! Face shields are so important!
Absolutely! We are happy to be able to contribute our small part.
How are your telepresence rollouts going in the current COVID crisis?
We have been prioritizing orders from hospitals and nursing homes. We delivered 26 robots to Mt. Sinai hospital in New York City recently. They need a way for nurses to connect with patients remotely and only see them when they absolutely need to. The ICU department is also using Ohmni to connect patients with their family. It’s been encouraging because these patients need a lot of support and positive energy.
How do you see the economic impact of COVID impacting OhmniLabs in the long term?
We are incredibly lucky to have a product that is bringing tremendous value during this difficult time. In the long term, we will see more demand for medical applications as an extension of telepresence. For example, we have been exploring how to add medical trays and UVC sanitation devices to our robots.
I would love to hear more about your rollout process. We have 81 robots to deliver and I’m putting together the production plan now. ANY tips, warnings, etc. would be greatly appreciated.
Certainly! We built an enterprise portal and gave access to hospital administrators. This portal allows them to share access, limit access when necessary, keep track of the robot, and streamline operations. This enterprise portal has been instrumental for our successful deployments.
What was your “aha moment” being in the Robotics industry?
For us it’s when we see how our customers use robots as extensions of themselves. Many suffer from trauma and are now using our robots to regain their normal life. They are attending class, visiting friends, and carrying out their normal life. It really helps alleviate the pain of whatever trauma they have to endure.
Just now we received notes from Phoenix Hospital where they use Ohmni in their Child Life Zone describing all of the ways their kids are using the robots:
- Patients can be guest hosts on CLZTV through a robot.
- Guest hosts from their community (such as sports teams) “beam into” the robots when provided access through the share feature.
- An awesome artist “beamed into” a robot to draw caricatures for patients. The kids were thrilled to show him some of their own drawings, too!
- Patients visit The Child Life Zone play area virtually to participate in scavenger hunts, find Easter eggs, complete obstacle courses, explore 3D printing, use the photo booth, race, and play games like Giant Connect 4 and Hide-n-Seek.
It’s heartwarming and also a big aha moment for us.
What adjustments beyond adapting to add medical trays, etc. have you had to make to your roadmap due to COVID? And how do you see upcoming technologies like 5G impacting your plans and potentially opening up new customers/opportunities?
Beyond the technical adaptations, we have been adjusting business and customer success strategies to align better with enterprise clients in the healthcare space. We’ve been testing 5G capabilities and I think it will really help with deployment, especially of outdoor robots! We are working on a very interesting custom robotics development project to debut later this year. I will share more info when it’s available!
What were some of the challenges you faced while building these robots?
Ah, so many challenges! I think we can all agree that hardware is just hard. Supply chain is a big one these days due to the global pandemic. Fortunately, we have been building our own 3D printers and the majority of our robots are 3D-printed. So the supply chain issues are tough but not a show-stopper. We continue to persevere because it’s fun and humbling to physically see people interacting with our robots.
Looking at your solution space, it looks like your technology covers everything from the technology required to build the robot to cloud services — it seems like complete e2e (end-to-end) coverage. What do you see as your successes in achieving this e2e coverage and how do you see it scaling to the future?
Thank you for that observation! You’re absolutely right — we do offer an end-to-end robotics solution. It is possible due to our additive manufacturing, modular tech library and robust cloud computing. Whenever we build new tech, we make sure that each component works on its own.
My team at OhmniLabs is known for bringing robotics concepts to life in record time. You may all be familiar with the recent ‘viral’ use case of 2 Avatar robots being used at the Tokyo BBT University graduation.
We took that robot concept and brought it to production in just 6 months. It required us to work on many processes simultaneously. Normally you design a component, test it and move to the next component, but we do about a dozen designs simultaneously. We push the design team to get to 80% satisfactory (quick turn) and then prototype using our own 3D printer farm. Once we put components together, we can test and fine tune a lot faster.
We also start the manufacturing and assembly process as we’re validating the design. Once the engineering is sound, the production team is ready to manufacture the robots.
In the future, we will scale by continuing to add to our library tech stack through these highly customized projects, and of course manufacturing awesome custom robots like Avatar robots.
Do you have AI integrated into your product?
Yes, we do! Ohmni uses AI in the autodock feature, recognizing users and interacting as a receptionist, wake-up words, and autonomy (which is in the works!)
I would love to hear about Kambria — what was the inspiration?
Kambria was born out of our desire to push for open innovation. Through the Kambria platform, we hope to incentivize developers to work together and not have to invent the wheel every time they have a great product idea. A lot of specialized robots can be built on a general platform.
We have shared a lot of AI and robotics tech with our Kambria community and hope to push AI-based applications in robotics. For example, we just ran a competition to categorize doorknobs. It sounds silly but recognition is the first step to actually opening that door!
Your profile states that you are once in Academia. What mistakes do you see as a trend amongst your students?
Yes, I was a professor at NYU for 6+ years. Given the readily available wealth of information out there, I think the teacher’s job has changed to teaching students how to think critically, how to make sense of all that information and then work on its application. The biggest mistake I see is related to the students’ focus on text-book knowledge and not on applications. Also, it’s a mistake to not network! No matter how smart you are, you won’t be recognized if you don’t make connections and share ideas.
How was your experience of moving from academia to a robotics startup and what led to the shift?
The underlying drive for me to pursue a career in academia and then a startup is my desire to learn something new. I am learning a lot every day operating a robotics startup, perhaps even more so than when I was a professor at NYU. It’s a hyper fast-paced environment, always changing.
Academia, on the other hand, is much slower. Many research projects span 3-5 years, but the vision is long-term. The basis of academic research is to start with: what will the world look like 10-30 years from now? In a startup, with every challenge thrown at us daily, we are presented with the opportunity to adapt and come out stronger. The vision may be long-term but the execution is constantly changing in the short-term. My experience has been good overall, but the pace did take some getting used to.
Nowadays I enjoy the changing execution. I see a ton of value in long-term research, however, and I hope the two fields can collaborate more closely. Perhaps our will be the big springboard for robotics HRI (human-robot interaction) research!
Until we fully understand how people interact with robots in various environments, our products will continue to be capped by low adoption.
On that note, what is your vision of robotics in the coming years?
We predict robotics will be dominated by service robots — robots that work alongside people to enhance their lives. Whether in a home, office, or hospitals, service robots can help people be more productive. There are some major technological hurdles our industry has to solve. The silver lining that this pandemic has created a paradigm shift in people’s acceptance of robots.
Yes, very true! Suddenly the need for delivery robots has also been understood. What are your thoughts on the self-driving car industry?
Oh my goodness – this is definitely a personal interest since my background is in Transportation Engineering. And now I get to see the technology side. On one hand, the promise of life saving is so great with autonomous cars. On the other hand, people will not tolerate any mistakes here. Tesla does their rollouts slowly hoping to change the public’s mindset along the way. It’s working somewhat.
Full autonomy is a long way off, not only because of sensor technology but also due to policy and implementation issues. I believe that autonomous driving will be great for society ONLY IF the policies are done well to regulate its use in order to truly realize the solutions in solving fundamental environmental and societal risks.
I have a background in motion planning. How important is that aspect in the robots OhmniLabs builds? Is it RL-based or just algorithmic?
We are implementing both. For simple tasks like autodocking, it’s just algorithmic, but we use reinforcement learning in more complex autonomy.
What is autodocking?
It’s a feature that allows our Ohmni robot to find its charging dock and dock itself. This is great from a user’s perspective. They just click a button and the bot docks itself.
Thank you, Tra for your time. It was great to learn more about you and OhmniLabs! Thank you for sending out those 26 robots and the face shields! It was great and so much fun!
Thank YOU, Prenali, for facilitating. It’s great to chat with everyone. I hope we can connect more in the future. Feel free to reach out to me anytime Please do take care and stay safe!
[The End]
A huge thanks to Silicon Valley Robotics and Women in Robotics for featuring Tra in this AMA. If you’d like Tra any questions about robotics, please contact us on Twitter.
About OhmniLabs
OhmniLabs is changing the paradigm of robotics development by creating flexible robotics modules and scalable additive manufacturing processes. Our flagship product, the Ohmni® Robot, is an award-winning telepresence robot that transforms how people connect, from their homes, businesses, classrooms, to hospitals. Our Ohmni® Developer Edition is an affordable, human-scale, open platform that includes our cloud-based JS framework OhmniAPI. We offer developers, researchers, and students the ability to program Ohmni® robots as easily as developing web pages.
About Silicon Valley Robotics
Silicon Valley Robotics is a not-for-profit coalition of robotics companies and startups, representing the robotics cluster of northern California. Silicon Valley Robotics was launched in 2010 by a network of robotics companies in the greater Bay Area including current sponsoring members, Fetch Robotics, Festo, EandM Engineering, SICK Sensor Intelligence, SRI International, Cruise Automation and Harmonic Drive.