Arun Narenthiran Sivakumar: CropFollow++: Robust under-canopy navigation with keypoints

Bionote

Arun Narenthiran Sivakumar is a Ph.D. candidate in Computer Science at the University of Illinois Urbana-Champaign, advised by Prof. Girish Chowdhary. His research focuses on developing robust visual navigation systems for mobile robots operating in challenging under-canopy environments. Broadly, his work lies at the intersection of field robotics, robot perception, and robot learning. His work has been recognized as ‘Outstanding Demo Paper Award Finalist’ at RSS 2024. Before his Ph.D., he earned an M.S. in Agricultural and Biological Systems Engineering with a minor in Computer Science from the University of Nebraska-Lincoln and a B.Tech. in Mechanical Engineering from VIT University, Vellore, India.

We present a robust vision-based navigation system for under-canopy agricultural robots using semantic keypoints. Autonomous under-canopy navigation is challenging due to the tight spacing between the crop rows (~0.75 m) and the degradation in RTK-GPS accuracy from multipath error. Our system, CropFollow++, introduces modular perception architecture with a learned semantic keypoint representation. This semantic keypoint representation is more modular, and more interpretable than the earlier work called CropFollow, and provides a confidence measure to detect occlusions. CropFollow++ considerably outperformed CropFollow in terms of the number of collisions (13 vs. 33) in field tests spanning ~1.9km each in challenging late-season fields with significant occlusions. We also deployed CropFollow++ in multiple under-canopy cover crop planting robots on a large scale (25 km in total) in various field conditions and we discuss the various failure modes observed.

Co-Authors: Mateus Valverde Gasparino, Michael McGuire, Vitor Akihiro Hisano Higuti, M. Ugur Akcal, Ghirish Chowdhary