<link rel="stylesheet" href="styles.f3b1fba60ec7970c.css">

IoT-Based Environmental Stress Index and LLM Diagnostics for Laying Quail Egg Production during Pancaroba in Siwatu Village, Batang Regency, Central Java

dc.contributor.authorSupriyanto
dc.date.accessioned2026-11-12T06:15:02Z
dc.date.issued2026
dc.description.abstractLaying quail (Coturnix japonica) production in Batang Regency, Central Java faces recurring stress during the pancaroba season the wet-to-dry transitional period marked by unpredictable temperature swings, persistently high humidity, and ammonia accumulation inside cages. Small-scale farmers in the region rely on visual inspection and smell to assess cage conditions, meaning that production drops are discovered after the fact rather than anticipated. This research designed, built, and tested an integrated monitoring system combining Internet of Things sensors, a composite Environmental Stress Index (ESI), and a Claude Sonnet Large Language Model (LLM) diagnostic layer, deployed at a working laying quail farm in Siwatu Village, Wonotunggal District, Batang Regency, Central Java. The system used an ESP32 microcontroller connected to temperature, humidity, ammonia, and light intensity sensors that transmitted readings every 30 minutes to a cloud backend. Each night the backend computed a daily ESI score a weighted composite of thermal load, ammonia concentration, and light adequacy classified the result into Good, Needs Attention, or Critical, then sent the summary to Claude Sonnet to generate a structured diagnosis covering cage status, identified problems, and recommended actions. Results were delivered to the farmer through a mobile application on his Android phone. Field deployment across 30 valid days covering approximately 1,800 laying quail showed that environmental conditions throughout the pancaroba window stayed consistently above safe thresholds, with the daily ESI averaging 0.53 and not a single day qualifying as Good 83.3% of days fell in the Needs Attention band and 16.7% in Critical. Statistical analysis produced a significant negative correlation between the daily ESI and daily egg production (r = −0.361, p = 0.050), confirming that the composite index tracks laying performance under real field conditions. The LLM diagnostic layer achieved 100% sensitivity for Critical days and consistently erred toward greater caution rather than less, a pattern clinically safer than conventional threshold-only monitoring. A documented intervention episode near the end of the observation period showed a 35% reduction in ammonia concentration following an Emergency-level alert, providing direct field evidence that the system prompted timely corrective action. All seven functional test scenarios passed.
dc.identifier.urihttps://repository.uii.ac.id/handle/123456789/67590
dc.language.isoen
dc.publisherUniversitas Islam Indonesia
dc.subjectInternet of Things
dc.subjectEnvironmental Stress Index
dc.subjectLarge Language Model
dc.subjectlaying quail
dc.subjectpancaroba season
dc.subjectegg production
dc.subjectBatang Regency
dc.titleIoT-Based Environmental Stress Index and LLM Diagnostics for Laying Quail Egg Production during Pancaroba in Siwatu Village, Batang Regency, Central Java
dc.typeThesis

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
22523294.pdf
Size:
7.35 MB
Format:
Adobe Portable Document Format