Aging Fault Detection and Failure Prediction Technologies within IoT Edge Devices

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K20234
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 60040:Computer system-related
• Research Institution: Ehime University
• Principal Investigator: Wang Senling 愛媛大学, 理工学研究科(工学系), 講師 (90735581)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 論理再構成デバイス / 信頼性設計 / テスト / 劣化検知

Outline of Final Research Achievements

In this research, we have established techniques to enhance the reliability throughout the lifecycle - from manufacturing to operation, to retirement - of Memory-based Programmable Devices (MPD), which are being developed as next-generation edge devices. Our primary accomplishments are as follows: ① To improve the manufacturing yield of MPD devices, we have established a high-quality test generation method for detecting and locating faults in the interconnecting wiring between memory cells.② We proposed a multi-cycle power-on self-test to enhance the reliability of devices during operation.③ We have proposed an aging detection technique for MPD devices in operation (oscillation-count integrated reconfigurable delay measurement circuit), and have established a method for its implementation in MPD devices.

Free Research Field

LSIテスト容易化設計

Academic Significance and Societal Importance of the Research Achievements

IoT技術の普及により、高性能・高信頼性のエッジデバイスが必要となっている。次世代エッジデバイスとして開発が進むMPD（Memory-based Programable Device）は、稼働中の異常に対する「予防」、「検知」、「回復」の信頼性要件を満たさなければならない。本研究では、MPDの製造から運用、リタイアまでのライフサイクル全体での高信頼性を向上させる技術を確立し、激化する次世代エッジデバイスの研究開発競争において、日本発の信頼性の高いMPD技術の普及を加速することに貢献する。