AI project promises better agri yield; concept tested in two farmer co-ops

Bengaluru: A new, AI-driven approach to crop planning and agri input procurement is promising to enable farmers with informed choices and higher revenues.

GRAMA (Game Theory, Random Processes, Artificial Intelligence, and Machine Learning for Agriculture) is a NABARD-backed project that works on the principles of deep learning and game theory.

The project is led by Y Narahari, honorary professor at the IISc’s Department of Computer Science and Automation, and involves experts from the University of Agricultural Sciences, GKVK.

GRAMA is aimed at maximising farmers’ revenue through crop planning, crop recommendation, procurement mechanisms, market mechanisms, and price prediction.

The project uses existing data on the types of soil, district-specific crop variety, demand, yield, historical pricing, seasonal variations, and sociocultural factors. Researchers projected a 60 per cent rise in the farmers’ income if they applied the right variations in the combinations in crops.

“It means that we are not even planning the crops properly. People who have grown a particular variety of crop over a long time may not want to shift to a new crop because they believe they already have the required expertise (and are averse to change),” Prof Narahari told DH.

The model has a provision to find out how the revenue changes with alterations in the crop combinations. “Even with 50 per cent compliance, the revenue can be 10 per cent higher,” he said.

The researchers see the empowerment of Farmer Producer Organisations (FPOs) as crucial in scaling up the project. Earlier this year, they initiated pilot projects at FPOs in Anekal and Rajaghatta where the FPOs showed much enthusiasm though farmers needed to be convinced about the technology. However, Prof Narahari believes that a combination of government mandate and greater involvement of progressive farmers can, over time, make GRAMA integral to farming practices.

A pilot of the procurement mechanism was tested at the two FPOs through a mobile app developed by Bharat Electronics Limited (BEL). Farmers’ cooperatives collect the requirements from the farmers. Based on these requirements and existing information on multiple factors including weather and soil type, the app recommends aggregate amounts of seeds, fertilisers, and pesticides to be procured. The model proposes a system that guarantees the farmer procurement at an affordable cost and reliable quality.

Prof Narahari said the adoption of GRAMA needed to be backed with efforts to minimise transportation and optimise the utilisation of available warehouses and other resources. He also pointed to many practical challenges.

Agricultural economist Prakash Kammaradi explained that it is crucial for any futuristic agricultural invention to ensure a more accurate remunerative price discovery, aggregation of produce, and effective distribution mechanisms.

“Agrarian issues are complex and challenging. The use of advanced technology can give us a deep understanding of these challenges, along with field experience. But relying entirely on it for solutions may not be sustainable,” he said.