LP-OPTIMA: A Framework for Prescriptive Maintenance and Optimization of IoT Resources for Low-Power Embedded Systems.

Area of Science:

• Computer Engineering
• Embedded Systems
• Artificial Intelligence

Background:

• Low-power embedded systems are crucial for data collection and exchange, but face maintenance challenges due to fault prediction and monitoring limitations.
• Existing resource management frameworks lack specific protocols for low-power embedded system failures.
• Identifying failures and implementing customized reaction mechanisms remain complex.

Purpose of the Study:

• To address the maintenance and fault detection gaps in low-power embedded systems.
• To propose a trilateral framework including stakeholder guidance, automated control, and AI-driven malfunction detection.
• To enhance the reliability and fault tolerance of these systems.

Main Methods:

• Development of a trilateral framework with periodic stakeholder prescriptions and automated control mechanisms.
• Implementation of a backup AI malfunction detection module.
• Design and development of three novel autonomous embedded systems using ARM Cortex cores for evaluation.

Main Results:

• The AI malfunction detection module demonstrated outstanding performance in real-life testing.
• Evaluation metrics consistently exceeded 98% across the developed embedded systems.
• The approach proved effective in enhancing fault tolerance and maintenance capabilities.

Conclusions:

• The proposed AI malfunction detection module significantly improves the reliability of low-power embedded systems.
• The trilateral framework offers a comprehensive solution for system maintenance and failure prevention.
• The developed approach is highly effective and reliable for critical embedded applications.