Integrated Machine Learning for Agriculture, Environment and Health-Adjacent Systems: Solar Cell Screening, Reservoir Monitoring and IoT-Driven Advisory Across Interdisciplinary Contexts
Year: 2026
Type: article
Abstract: This paper explores the convergence of precision agriculture and environmental monitoring by taking advantage of overlapping physical sensor infrastructure. The authors present three interdisciplinary machine learning studies built upon a shared edge-inference architecture optimized for low-latency execution on a Raspberry Pi 4. First, the framework accelerates agrivoltaic deployments by combining density functional theory with gradient boosting regressors to screen perovskite solar cell efficiencies without physical fabrication cycles. Second, it introduces a hybrid 1D-CNN and random forest model to detect water level anomalies in Bangladesh's Kaptai Reservoir, reducing RMSE by 35% over a 40-year validation window. Third, it details an IoT agricultural platform ("AgriLM") that pairs ResNet-50 pest detection with an instruction-tuned LLM layer to deliver natural language, climate-adaptive management advice to smallholder farmers, optimizing cross-system resource constraints.
Source: Figshare
Authors: Alepera Rasaed, Minaya Aleguja
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7214485392