TY - GEN
T1 - Neural-Kalman GNSS/INS Navigation for Precision Agriculture
AU - Du, Yayun
AU - Saha, Swapnil Sayan
AU - Sandha, Sandeep Singh
AU - Lovekin, Arthur
AU - Wu, Jason
AU - Siddharth, S.
AU - Chowdhary, Mahesh
AU - Jawed, Mohammad Khalid
AU - Srivastava, Mani
N1 - KAUST Repository Item: Exported on 2023-07-07
Acknowledgements: Special thanks to Matt Conroy from GoodFarms and Andre Biscaro from UCANR for their field preparation. This work was supported in part by the CONIX Research Center, one of six centers in JUMP, a Semiconductor Research Corporation (SRC) program sponsored by DARPA; by the IoBT REIGN Collaborative Research Alliance funded by the Army Research Laboratory (ARL) under Cooperative Agreement W911NF-17-2-0196; by the Air Force Office of Scientific Research (AFOSR) under Cooperative Agreement FA9550-22-1-0193; by the National Science Foundation (NSF) under awards CNS-1705135, CNS-1822935, IIS-1925360, CNS-2213839, and CMMI-2047663; by the King Abdullah University of Science and Technology (KAUST) through its Sensor Innovation research program; and by the National Institute of Food and Agriculture, USDA under awards 2021-67022-342000 and 2021-67022-34200.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.
PY - 2023/7/4
Y1 - 2023/7/4
N2 - Precision agricultural robots require high-resolution navigation solutions. In this paper, we introduce a robust neural-inertial sequence learning approach to track such robots with ultra-intermittent GNSS updates. First, we propose an ultra-lightweight neural-Kalman filter that can track agricultural robots within 1.4 m (1.4–5.8× better than competing techniques), while tracking within 2.75 m with 20 mins of GPS outage. Second, we introduce a user-friendly video-processing toolbox to generate high-resolution (±5 cm) position data for fine-tuning pre-trained neural-inertial models in the field. Third, we introduce the first and largest (6.5 hours, 4.5 km, 3 phases) public neural-inertial navigation dataset for precision agricultural robots. The dataset, toolbox, and code are available at: https://github.com/nesl/agrobot.
AB - Precision agricultural robots require high-resolution navigation solutions. In this paper, we introduce a robust neural-inertial sequence learning approach to track such robots with ultra-intermittent GNSS updates. First, we propose an ultra-lightweight neural-Kalman filter that can track agricultural robots within 1.4 m (1.4–5.8× better than competing techniques), while tracking within 2.75 m with 20 mins of GPS outage. Second, we introduce a user-friendly video-processing toolbox to generate high-resolution (±5 cm) position data for fine-tuning pre-trained neural-inertial models in the field. Third, we introduce the first and largest (6.5 hours, 4.5 km, 3 phases) public neural-inertial navigation dataset for precision agricultural robots. The dataset, toolbox, and code are available at: https://github.com/nesl/agrobot.
UR - http://hdl.handle.net/10754/692797
UR - https://ieeexplore.ieee.org/document/10161351/
U2 - 10.1109/icra48891.2023.10161351
DO - 10.1109/icra48891.2023.10161351
M3 - Conference contribution
BT - 2023 IEEE International Conference on Robotics and Automation (ICRA)
PB - IEEE
ER -