Resilient and Modular Subterranean Exploration with a Team of Roving and Flying Robots
Sebastian Scherer,Vasu Agrawal,Graeme Best,Chao Cao,Katarina Cujic,Ryan Darnley,Robert DeBortoli,Eric Dexheimer,Bill Drozd,Rohit Garg,Ian Higgins,John Keller,David Kohanbash,Lucas Nogueira,Roshan Pradhan,Michael Tatum,Vaibhav Viswanathan,Steven Willits,Shibo Zhao,Hongbiao Zhu,Dan Abad,Tim Angert,Greg Armstrong,Ralph Boirum,Adwait Dongare,Matthew Dworman,Shengjie Hu,Joshua Jaekel,Ran Ji,Alice Lai,Yu Hsuan Lee,Anh Luong,Joshua Mangelson,Jay Maier,James Picard,Kevin Pluckter,Andrew Saba,Manish Saroya,Emily Scheide,Nathaniel Shoemaker-Trejo,Joshua Spisak,Jim Teza,Fan Yang,Andrew Wilson,Henry Zhang,Howie Choset,Michael Kaess,Anthony Rowe,Sanjiv Singh,Ji Zhang,Geoffrey Hollinger,Matthew Travers
DOI: https://doi.org/10.55417/fr.2022023
2022-03-10
Field Robotics
Abstract:Subterranean robot exploration is difficult, with many mobility, communications, and navigation challenges that require an approach with a diverse set of systems, and reliable autonomy. While prior work has demonstrated partial successes in addressing the problem, here we convey a comprehensive approach to address the problem of subterranean exploration in a wide range of tunnel, urban, and cave environments. Our approach is driven by the themes of resiliency and modularity, and we show examples of how these themes influence the design of the different modules. In particular, we detail our approach to artifact detection, pose estimation, coordination, planning, control, and autonomy, and we discuss our performance in the tunnel, urban, and self-organized cave circuits of the DARPA Subterranean Challenge. Our approach led to a winning result in the tunnel circuit, and placing second in the urban circuit event. We convey lessons learned in designing and testing a resilient system for subterranean exploration that can generalize to a large range of operating conditions, and potential improvements for the future.