Yan Watequlis Syaifudin, Muhamad Syaroful Anam, Triana Fatmawati, Nobuo Funabiki, Noprianto, Chandrasena Setiadi, Rokhimatul Wakhidah, Amar Alpabet Fuaduzakiah, M. Hasyim Ratsanjani
The increasing demand for efficient and scalable digital tools in agriculture has led to the development of modular mobile applications capable of dynamically adapting to user needs and environmental constraints. This paper presents a mobile smart farming application built using Dynamic Feature Modules (DFMs), enabling on-demand feature delivery to reduce initial download size, optimize resource usage, and enhance performance. The system integrates IoT-based environmental monitoring, AI-driven plant disease detection, and market analytics through a modular architecture that supports seamless feature expansion. Performance evaluation was conducted under controlled 4 G network conditions (15 Mbps average throughput) on mid-range Android devices, measuring execution latency, memory footprint, and module delivery efficiency. The technical evaluation shows that the base module size was reduced to 15.1 MB from 48.7 MB in a monolithic build, with average load times of 1.2 seconds for the core module and 1.1 to 2.3 seconds for on-demand modules. Stability tests confirmed consistent CPU and memory usage, while backend API response times averaged 0.8 seconds. User engagement also improved due to contextual feature prompting and personalized access. These results demonstrate that DFM-based modularization enhances scalability, maintainability, and user experience for future smart agriculture platforms. © 2025 IEEE.
State Polytechnic of Malang, Dept. of Information Technology, Malang, Indonesia; Okayama University, Dept. of Elec. and Comm. Engineering, Okayama, Japan; PT Link Apisindo Media, Dept. of Information Technology, Malang, Indonesia