DOI QR코드

DOI QR Code

LSTM-based Fire and Odor Prediction Model for Edge System

엣지 시스템을 위한 LSTM 기반 화재 및 악취 예측 모델

  • Received : 2022.01.04
  • Accepted : 2022.01.10
  • Published : 2022.02.28

Abstract

Recently, various intelligent application services using artificial intelligence are being actively developed. In particular, research on artificial intelligence-based real-time prediction services is being actively conducted in the manufacturing industry, and the demand for artificial intelligence services that can detect and predict fire and odors is very high. However, most of the existing detection and prediction systems do not predict the occurrence of fires and odors, but rather provide detection services after occurrence. This is because AI-based prediction service technology is not applied in existing systems. In addition, fire prediction, odor detection and odor level prediction services are services with ultra-low delay characteristics. Therefore, in order to provide ultra-low-latency prediction service, edge computing technology is combined with artificial intelligence models, so that faster inference results can be applied to the field faster than the cloud is being developed. Therefore, in this paper, we propose an LSTM algorithm-based learning model that can be used for fire prediction and odor detection/prediction, which are most required in the manufacturing industry. In addition, the proposed learning model is designed to be implemented in edge devices, and it is proposed to receive real-time sensor data from the IoT terminal and apply this data to the inference model to predict fire and odor conditions in real time. The proposed model evaluated the prediction accuracy of the learning model through three performance indicators, and the evaluation result showed an average performance of over 90%.

최근 인공지능을 활용한 다양한 지능형 응용서비스 개발이 활발히 진행 중이다. 특히, 제조 산업 현장에서는 인공지능 기반 실시간 예측서비스 연구가 활발히 진행 중이며 이중 화재 및 악취를 감지·예측할 수 있는 인공지능 서비스에 대한 요구가 매우 높다. 하지만 기존 감지·예측시스템은 화재 및 악취 발생 예측이 아닌 발생 후 감지 서비스가 대부분이다. 이는 인공지능 기반 예측서비스 기술이 적용되어 있지 않기 때문이다. 또한, 화재 예측 및 악취 감지·예측서비스는 초저지연 특징을 가진 서비스이다. 따라서 초저지연 예측서비스를 제공하기 위해 엣지 컴퓨팅 기술이 인공지능 모델과 결합되어 클라우드에 비해 빠른 추론 결과를 현장에 빠르게 적용할 수 있도록 개발 중이다. 따라서 본 논문에서는 제조 산업 현장에서 가장 많이 요구되는 화재 예측 및 악취 감지·예측에 사용할 수 있는 LSTM 알고리즘 기반 학습모델을 제안한다. 또한, 제안하는 학습모델은 엣지 다바이스에 구현이 가능하도록 설계하였으며 사물인터넷 단말로부터 실시간 센서데이터를 수신하고 이 데이터를 추론 모델에 적용하여 화재 및 악취 상태를 실시간으로 예측할 수 있도록 제안한다. 제안된 모델은 3가지 성능 지표를 통해 학습모델의 예측 정확도를 평가하였으며 평가 결과는 평균 90% 이상 성능을 보였다.

Keywords

Acknowledgement

이 논문은 IITP/NIST SW 컴퓨팅산업원천기술개발과제(20200001160012007)와 2021학년도 동의대학교 교내연구비(202101850001)에 의해 연구되었음.

References

  1. M. Nakip and C. Guzelis, ''Development of a multi-sensor fire detector based on machine learning models,'' in 2019 Innovations in Intelligent Systems and Applications Conference (ASYU), pp.1-6, Oct. 2019.
  2. L. Wu, L. Chen, and X. Hao, ''Multi-sensor data fusion algorithm for indoor fire early warning based on BP neural network,'' Information, Vol.12, No.2, pp.59, Jan. 2021. https://doi.org/10.3390/info12020059
  3. A. Sharma, P. K. Singh, and Y. Kumar, ''An integrated fire detection system using IoT and image processing technique for smart cities,'' Sustainable Cities and Society, Vol.61, Art. No.102332, Oct. 2020.
  4. Y. Mao, C. You, J. Zhang, K. Huang, and K B. Letaief, "A survey on mobile edge computing: The communication perspective," IEEE Communications Surveys Tutorials, Vol.19, No.4, pp.2322-2358, 2017. https://doi.org/10.1109/COMST.2017.2745201
  5. R. Chanonsirivorakul and N. Nimsuk, "A study of relationship between sensor response and odor perception in human," in International Electrical Engineering Congress (iEECON), pp.1-4, Mar. 2018.
  6. D. Wu, D. Luo, K. Wong, and K. Hung "POP-CNN: Predicting odor pleasantness with convolutional neural network," IEEE Sensors Journal, Vol.19, No.23, pp.11337-11345, 2019. https://doi.org/10.1109/jsen.2019.2933692
  7. Z. Xu, Y. Guo, and J. H. Saleh, "Advances toward the next generation fire detection: Deep LSTM variational auto-encoder for improved sensitivity and reliability," IEEE Access, Vol.9, pp.30636-30653, 2021. https://doi.org/10.1109/ACCESS.2021.3060338