JPL team co-led by Luca Carlone participates in DARPA Subterranean Challenge

Image Credit

Main image from NASA/JPL-Caltech. Team photo below provided by DARPA.

Image Caption

A new generation of robots can autonomously explore caves, pits, tunnels and other subsurface terrain.

Article Author

August 14, 2019

Updated August 29, 2019: We are happy to share that the CoSTAR team took second place in the Tunnel Circuit competition of the DARPA Subterranean Challenge on Aug. 22, 2019. The Subterranean Challenge will continue with the Urban Circuit in February 2020.


LIDS faculty member Luca Carlone is the MIT lead of Team CoSTAR — a multi-university team led by NASA’s Jet Propulsion Laboratory (JPL) that is participating in the Defense Advanced Research Projects Agency (DARPA) Subterranean Challenge Systems Competition. 

As the JPL press release explains, “the [DARPA] competition is intended to develop technology for first responders and the military to map, navigate and search underground. But technology developed for the competition will also lay the foundation for future NASA missions to caves and lava tubes on other planets.”

The first phase of the competition, the Tunnel Circuit, is being held August 15 - 22 in Pittsburgh, PA, in formerly operational mines managed by the National Institute for Occupational Safety and Health Mining Program. In this circuit, teams use autonomous robots to map and detect objects in a portion of the tunnels.



The Collaborative SubTerranean Autonomous Resilient Robots Systems team (CoSTAR) is one of 11 teams competing in the challenge, and is a collaboration between JPL, MIT, Caltech, KAIST, LTU and several industry partners. (In addition to Prof. Carlone, MIT is represented by LIDS student Yun Chang. Former LIDS postdoc Ali Agha is the Principal Investigator on the project.) The team will use wheeled rovers and tank-like tracked robots to cover the ground, and flying drones to find out-of-reach objects.

To learn more about Team CoSTAR see:

To learn more about the DARPA SubT Challenge see: