Applying AI to Advance Space Exploration – Intel Chip Chat – Episode 627
In this Chip Chat podcast, Zahi Kakish, a doctoral student focusing on swarm robotics at Arizona State University, joins Shashi Jain, innovation manager at Intel, to talk about some remarkable work that’s emerged from NASA FDL, the space agency’s Frontier Development Lab.
Designed as an eight-week challenge, FDL is an AI accelerator project set up by NASA Ames, the SETI Institute, and private partners. Its goal? To bring together the best minds in AI and planetary science to tackle challenges facing NASA and the commercial space industry.
This past summer, with support from Intel principal engineers and an Intel Xeon processor-based server at SETI, Kakish and his FDL team created a planning tool called the Mission Planner for Cooperative Multi-Agent Systems, MARMOT for short. Through the use of IA and machine learning, it enables two semi-autonomous rovers to communicate and work together to solve a task.
With its collaborative systems and assistive AI, MARMOT delivers a significant performance improvement over missions that employ a single robot with a human operator. It represents a fresh and highly optimized approach to autonomy, and it’s available as open source code.
To access the code for all NASA FDL projects, visit:
gitlab.com/frontierdevelopmentlab
Additional information is available on the FDL site at:
frontierdevelopmentlab.org
To learn about the Autonomous Collective Systems Lab at Arizona State University and their work on swarm robotics, go to http://faculty.engineering.asu.edu/acs/.
More about Intel’s involvement in FDL is posted on:
intel.com/ai
To access Intel AI DevCloud, a cloud hosted hardware and software platform that helps developers, researchers, and startups get started on AI projects, visit:
ai.intel.com/devcloud
Posted in:
Artificial Intelligence, Audio Podcast, Intel, Intel Chip Chat