[Poster Presentation]MEMS Gas Sensor Array Fault Diagnostic Unit for Microsystem Applications - International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence ((ICSMD2024))

MEMS Gas Sensor Array Fault Diagnostic Unit for Microsystem Applications

ID：153 Submission ID：109 View Protection：ATTENDEE Updated Time：2024-10-23 10:02:35 Hits：38 Poster Presentation

Start Time：Pending （Asia/Shanghai）

Duration：Pending

Session：[No Session] » [No Session Block]

No files

Abstract

With the rapid development of the Internet of Things (IoT) and environment sensing systems, environment sensing intelligent microsystems based on microelectrome- chanical systems (MEMS) sensor arrays have vast application prospects and advantages in the fields of environmental monitoring, medical diagnosis, and battlefield sensing. The machine olfaction-oriented environment sensing intelligent microsystem is a microsystem that combines MEMS gas sensor array and pattern recognition algorithms to monitor the gases in the environment in real time, and its working performance is seriously dependent on the accuracy of the measurement signals from the MEMS gas sensor array. The performance of the microsystem relies heavily on the accuracy of the measurement signals from the MEMS gas sensor array. Due to the characteristics of the sensitive materials and the integrated fabrication process, factors such as environmental changes may cause failures of the MEMS gas sensor array, resulting in abnormal operation of the entire microsystem. Therefore, this paper designs a MEMS gas sensor array fault diagnostic unit for microsystem applications with the size of 19mm*19mm*3mm, which can identify five kinds of sensor faults, namely, shock fault, bias fault, constant output fault, power-down fault, and accuracy degradation fault, and has fault isolation and fault localisation functions, which can provide the basis for the recovery of faulty data of MEMS gas sensors. The multi-task fault diagnosis algorithm based on 1D CNN -LSTM on the fault diagnosis unit has a fault recognition accuracy of 99.93%, a fault isolation accuracy of 99.43%, and a fault localisation accuracy of 98.12%, occupies 194k of FLASH and 20k of SRAM, and has an average running time of 234ms, and runs with a power consumption of 11.58mW.

Keywords

MEMS gas sensor array,microsystems,fault diagnosis,multitask learning,embedded deployment

Speaker

Submission Author

FuJie Harbin Institute of Technology；Department of Measurement and Control Engineering at the School of Electronics and Information Engineering

YangJian No. 703 Research Institute of CSSC

FuHongshuo Harbin Institute of Technology；Department of Measurement and Control Engineering at the School of Electronics and Information Engineering

LiuBing Harbin Institute of Technology；Department of Measurement and Control Engineering at the School of Electronics and Information Engineering

ZhengWenbin Harbin Institute of Technology；School of Electronics and Information Engineering； Harbin 150080； P.R. China

FuPing Harbin Institute of Technology；Department of Measurement and Control Engineering at the School of Electronics and Information Engineering

Comment submit

All comments

Registration Paper Submission

Important Dates

~~15th August~~ ~~2024~~ 31st August 2024- Manuscript Submission

15th September 2024 - Acceptance Notification

1st October 2024 - Camera Ready Submission

1st October 2024 – Early Bird Registration

Contact Us

Website:

https://icsmd2024.aconf.org/