Background: Flooded rice ecosystems emit significantly more methane (CH₄) and nitrous oxide (N₂O) than other agricultural systems and, therefore, improving the efficiency of global rice production systems will be essential to meeting both food security and greenhouse gas emission reduction targets (especially CH₄ and N₂O) worldwide.
Objective: The purpose of this research is to create a carbon-smart rice production system that utilizes Edge-AI Autonomous Robots to implement site-specific management strategies for the Astragalus sinicus (Chinese milk vetch), a green manure crop that biologically fixes nitrogen, by the year 2050 by integrating technology into agriculture.
Methods: This method utilizes hyperspectral imaging, multispectral imaging, LiDAR canopy profiling, and electrochemical soil sensing with deep-learning models that are implemented on neuromorphic edge processors located on an autonomous ground robot. The simulations were done across 10,000 hectares of virtual fields to assess system performance.
Results: Increased soil organic carbon stock from A. sinicus efficiently maintained with robots using site specific intervention, reduced CH₄ emissions (compared to conventional methods) and decreased input of synthetic nitrogen by 12-19% and 35-52%, respectively.
Conclusion: Edge-AI robotics effectively tackle spatial heterogeneity in extensive rice systems, facilitating precision management that improves carbon sequestration and diminishes emissions.
Implications: The framework offers a scalable and technologically feasible approach for rice-based agriculture to achieve net-zero climate objectives while enhancing soil health, ecosystem services, and farm profitability by 2050.