AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0

October 2026
Vol-12, Issue-5
Paper ID: 29009
ISSN: 2395-4396
Downloads: 31
Download PDF

Abstract & Details

Research Area
Predictive Maintenance, Industrial IoT, Machine Learning, Explainable AI, Industry 4.0, Autonomous Decision Systems
Keywords
Predictive maintenance Random Forest Industry 4.0 Industrial Internet of Things large language models real-time monitoring event streaming priority scoring.
Abstract
Industrial maintenance has traditionally relied on reactive repair after breakdown or on fixed preventive schedules, both of which waste resources and still leave plants exposed to unplanned downtime. This paper presents the Autonomous Maintenance Decision Agent (AMDA), a real-time, AI-driven predictive maintenance platform that continuously ingests live sensor telemetry, forecasts impending equipment failures, and autonomously raises prioritized maintenance work orders with little human intervention. AMDA is built around four Random Forest classifiers, one for each of four industrial equipment types (CNC mill, CNC lathe, industrial pump, and assembly conveyor), each trained on a balanced synthetic dataset generated by a Node.js simulator and streamed to a Flask backend over Server-Sent Events (SSE). A feature-mapping layer translates the simulator’s generic telemetry into the equipment-specific features each model expects, and every prediction that clears a confidence threshold is enriched by a locally hosted large language model (Ollama running Qwen2.5) that turns a numeric failure label into a plain-language explanation, a likely root cause, and a recommended maintenance procedure. A priority-scoring engine combines failure severity, live sensor-derived risk, and machine criticality into a single score that drives work-order routing. Orders, schedules, and history are persisted in MongoDB, propagated through Kafka, and pushed to a React 19 dashboard over Socket.IO in real time. Evaluation on held-out data shows sub-second end-to-end inference latency and failure-classification accuracy above 92% across all four machine models. The result is a practical demonstration that Random Forest inference and local LLM reasoning can be combined into a single autonomous, explainable maintenance loop for Industry 4.0 settings.

Author Information

# Name Institute / Affiliation
1 Amisha lobo Alvas Institute of Engineering and Technology, Moodubidri
2 Santhosh gowda Alvas Institute of Engineering and Technology, Moodubidri
3 Keerthana SS Alvas Institute of Engineering and Technology, Moodubidri
4 Darshan Mutagond Alvas Institute of Engineering and Technology, Moodubidri
5 Jyotibha chinchankar Alvas Institute of Engineering and Technology, Moodubidri

How to Cite

Use the following formats to cite this article in your research.

APA Style
lobo, Amisha, gowda, Santhosh, SS, Keerthana, Mutagond, Darshan, & chinchankar, Jyotibha (2026). AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0. International Journal of Advance Research and Innovative Ideas In Education, 12(5), 857-862.
MLA Style
lobo, Amisha, et al. "AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0." International Journal of Advance Research and Innovative Ideas In Education, vol. 12, no. 5, 2026, pp. 857-862.
IEEE Style
Amisha lobo, Santhosh gowda, Keerthana SS, Darshan Mutagond, and Jyotibha chinchankar, "AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0," International Journal of Advance Research and Innovative Ideas In Education, vol. 12, no. 5, pp. 857-862, 2026.
Vancouver Style
lobo Amisha, gowda Santhosh, SS Keerthana, Mutagond Darshan, chinchankar Jyotibha. AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0. International Journal of Advance Research and Innovative Ideas In Education. 2026;12(5):857-862.
Harvard Style
lobo, Amisha, gowda, Santhosh, SS, Keerthana, Mutagond, Darshan, & chinchankar, Jyotibha (2026) 'AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0', International Journal of Advance Research and Innovative Ideas In Education, 12(5), pp. 857-862.
Chicago Style
lobo, Amisha, et al. "AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0." International Journal of Advance Research and Innovative Ideas In Education 12, no. 5 (2026): 857-862.
Turabian Style
lobo, Amisha, et al. "AMDA: Autonomous Maintenance Decision Agent—An AI-Based Predictive Maintenance Platform for Industry 4.0." International Journal of Advance Research and Innovative Ideas In Education 12, no. 5 (2026): 857-862.

Export Citation