AI-POWERED MOBILE ASSISTANT FOR PRECISION CINNAMON FARMING USING REAL-TIME WEATHER AND SOIL DATA IN THE SOUTHERN PART OF SRI LANKA

Abstract

Ceylon cinnamon (Cinnamomum verum) is a globally renowned agricultural product and a key contributor to Sri Lanka’s economy, particularly in the Southern Province. However, cinnamon farmers face persistent challenges due to climate variability, declining soil health, and limited access to real-time, localized agricultural insights, leading to suboptimal decision-making. To address these issues, this study presents the design, development, and evaluation of an Artificial Intelligence (AI)-powered mobile assistant aimed at enhancing precision cinnamon farming through the integration of real-time weather forecasts and soil condition data. The system was developed using the Flutter framework with Firebase backend services, incorporating RESTful Application Programming Interface (API)s for weather and soil data integration, while TensorFlow Lite enables on-device AI inference to ensure low-latency recommendations. Random Forest is used for predictive analytics, providing personalized, location-specific recommendations on irrigation scheduling, fertilization, plantation planning, and pest management. The assistant also supports multilingual interactions in Sinhala and Tamil, improving accessibility for rural farmers. Data were sourced from the Sri Lankan Department of Agriculture, meteorological services, public weather APIs, and scientific literature, complemented by field surveys and controlled experiments across three cinnamon cultivation sites over 6 months. The evaluation involved 20 AI-assisted farmers and a control group of 20 farmers following traditional practices. Field trials showed significant improvements: crop yields increased by 8-12%, water usage decreased by 15%, fertilizer usage reduced by 10%, and pesticide use lowered by 7% without compromising productivity. Over 80% of participants reported high satisfaction, citing usability and multilingual support as key enablers. This solution addresses gaps in existing agricultural platforms by delivering real-time, localized decision support tailored to the cinnamon sector. Future enhancements will integrate pest detection, disease prediction, and broader regional deployment, contributing to climate-resilient, sustainable, and data-driven agriculture in Sri Lanka.