Talking AI-Powered Robotics for Extreme Environments

Alicia Kavelaars is Co-Founder and Chief Technology Officer at OffWorld Inc. She brings over 15 years of experience in the aerospace industry developing and successfully launching systems for NASA, NOAA and the Telecommunications industry. In 2015, Alicia made the jump to New Space to work on cutting edge innovation programs. In her tenure at OffWorld, Alicia has led the development of AI based rugged robots that will be deployed in one of the most extreme environments on Earth as a precursor to swarm robotic space operations: deep underground mines. Alicia holds a MSc. and PhD from Stanford University and a BSc. in Theoretical Physics from UAM, Spain.


Namagiri Anand interacted with Alicia Kavelaars to get insights on AI-Powered Robots for extreme environment conditions, whether it be on earth, or in the space, including helping humans mine asteroids and develop settlements on the Moon and Mars.


​​

"We aim to develop a swarm robotic system that will enable the exploration and utilization of space, ranging from the harvesting of natural resources from asteroids, to building human settlements on the Moon and Mars. We are currently developing the first generation of an AI-powered robotic workforce for extreme environments on Earth."- Alicia Kavelaars, Offworld Inc. ​​

Namagiri: What is Offworld Inc.?


Alicia: OffWorld primary mission is to develop a swarm robotic system that will enable the exploration and utilization of space, ranging from the harvesting of natural resources from asteroids, the Moon or Mars, to building human settlements on the Moon and Mars and large infrastructure in free space. The main challenge of a swarm robotic system that can operate in space is its ability to work in extreme environments autonomously. We can only accomplish maximum autonomy through the use of Machine Learning and AI. Early on, OffWorld successfully bet on the idea that extreme environments on Earth can also greatly benefit from this technology, especially when putting humans out of harm’s way. This led to the inception of a dedicated program for terrestrial applications in extreme environments, starting with underground mining, and moving on to construction, disaster relief, repair and servicing of remote infrastructure. We are currently developing the first generation of an AI powered robotic workforce for extreme environments on Earth.


Namagiri: How did you embark on this role? And, what is your mandate?


Alicia: I am an aerospace engineer by training. I studied Theoretically Physics in Spain where I am originally from, and continued on to obtain Master’s and PhD Degrees in Aeronautics and Astronautics from Stanford University in the US. I was the typical young kid passionate about the Universe, with a deep interest in learning how it works, and how we can enable humans to explore it. I worked on NASA, NOAA, commercial Telecom, and other US Government programs for 15 years. In 2015, I made the jump to Space 2.0 to be part of the wave of entrepreneurs that is planning to revolutionize the way we see, access and utilize space. I co-founded and became the CTO of OffWorld in early 2016. My mandate is to ensure that we are developing our robotic workforce with the right technologies to ensure our robots will effectively operate autonomously in extreme environments here on Earth and in space. I constantly seek to strike a balance between developing new technologies where I find gaps in the state of the art, and ensuring we rapidly prototype our systems with the best available tech. AI is key for this goal, as there are many tasks that our robots will need to excel on that cannot really be performed with traditional hard coded algorithms.


Namagiri: How are you leveraging Artificial Intelligence and Machine Learning at Offworld Inc?


Alicia: We require our robots to react to unconstrained and unexpected conditions, and to be able to adapt based on trained knowledge that grows as the experience of our robot colony grows. Only AI can help us achieve this level of autonomy. We are utilizing various Machine Learning methodologies, in particular Deep Reinforcement Learning, for critical tasks that our robots need to perform equally or better than human experts. We believe that the use of AI should be made generic for any kind application, and towards that end we are building a framework that will allow us to obtain enormous amounts of robotic data to easily integrate new AI based functionality. We also believe however that AI is not a one shoe fits all kind of thing, you need to be quite selective in identifying those tasks that can really benefit from AI. There are some tasks that cannot really be performed without AI. Swarm collaboration can best be optimized with AI. And there are multiple applications that can be perfectly well implemented with traditional hard coded algorithms, were AI really does not provide a huge advantage. It is the proper blend of Machine Learning agents enhancing operations, while deployed in a traditional hard coded architecture, what gives our workforce the biggest bang for your buck.


Namagiri: Specifically, what barriers and challenges have you seen in adapting AI at Offworld?


Alicia: Not a barrier but a welcome challenge is the fact that we are really developing breakthrough technology. AI has evolved from pure data applications to robotic applications in structured environments. What we are trying to do is bring AI to the frontier of extreme unstructured environments. Our AI powered robotic workforce is going to work in extreme environments on Earth first: in deep underground mines, remote infrastructure projects, disaster relief areas, on the sea bed, and so on. And soon enough, they will operate in our ultimate destination off-world, to help humans mine asteroids and develop settlements on the Moon and Mars. Deploying this robotic workforce in these environments requires enhanced AI, and we find ourselves researching or defining and developing Machine Learning methodologies that will work best for us. We develop applications from the ground up, with detailed computer simulation followed by training in silica. As we go forward, we are developing ways in which we can minimize the amount of time it takes an agent to learn a new task, effectively reducing sample complexity. All of these activities and the development framework are new and have been built from the ground up, and we are very excited about how we can leverage this new technology moving forward.


Namagiri: When it comes to the skill-sets required for AI, how are you managing it within your company?


Alicia: We keep tabs of all the university and research institutions that are strong AI centers, in particular for the Machine Learning methodologies we are currently leveraging. We actively attend and participate in key conferences and have an on-going advisory relationship with key researchers in the field. This has helped us identify key talent that we have brought on-board at OffWorld in particular in the areas of deep reinforcement learning, embodied robotic intelligence, deep neural networks, robotic autonomy and planning, human-machine interaction and optimization theory. Our Machine Learning research engineering group has the mandate to research on specific topics that we have outlined in our Machine Learning roadmap with the goal of deployment and easy adaptation of our robotic workforce in the field. This research is supplemented by practical experiments and implementations of Machine Learning applications in silica. We have found this is a quite successful combination of long-term research and immediate actionable results that continuously validate our roadmap.


Namagiri: Let me ask one last question: What’s the best advice you’ve ever received?


Alicia: Be patient, be proactive, and build a strong Team. Patience is a very important skill to have all around, it will help you when there are setbacks in tech development, and dealing with your Team, customers, and suppliers. You need to be proactive when you see problems emerge and to stay the course while you solve them. Building and supporting a strong Team is in my opinion the most important thing any company, whether startup or established, must do. At the end of the day, you will only succeed in anything you are doing if you surround yourself with the best people, and you make sure you work to the best of each individual’s skills to fulfill your mission in order to achieve your vision.


  • LinkedIn Social Icon
  • Twitter Social Icon