IoT-Enabled Agricultural Knowledge Management System for Precision Farming: A Cloud-Edge Architecture for Smallholder Farmers in Indonesia

Abstract

Indonesia's agricultural sector, dominated by smallholder farmers managing plots of less than two hectares, faces mounting productivity challenges arising from climate variability, declining soil fertility, and inadequate access to timely, location-specific agricultural knowledge. Traditional extension service models are insufficient to provide the precision knowledge support required at the scale of Indonesia's 26 million smallholder farm households. Problem Statement: Existing agricultural information systems in Indonesia fail to capture real-time environmental conditions, integrate contextual sensor data with expert knowledge, and deliver actionable recommendations through channels accessible to smallholder farmers with limited digital literacy. Proposed System: This paper proposes IAKMS-Indo, an IoT-enabled Agricultural Knowledge Management System combining a distributed sensor network, cloud-edge processing architecture, and dynamic agricultural knowledge base to eliver precision farming recommendations. Methodology: IAKMS-Indo was evaluated using a simulated dataset of 580 farm monitoring episodes across rice and palm oil cultivation contexts representing three Indonesian provinces. Key Findings: The system achieved a crop stress detection accuracy of 88.6%, knowledge recommendation relevance precision of 84.2%, and average knowledge delivery latency of 94 ms. Applications: The system supports crop disease early warning, irrigation scheduling, fertilizer optimization, and market price knowledge dissemination.