ReFrAEN: a Reconfigurable Vibration Analysis Framework for Constrained Sensor Nodes

Thangarajan, Ashok Samraj; Yang, Fan; Joosen, Wouter; Michiels, Sam; Hughes, Danny

doi:10.1109/dcoss52077.2021.00033

Cited by 2 publications

(1 citation statement)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Acoustic Anomaly Detection (AAD) is the task of detecting abnormalities in machine operation by analyzing the characteristic of its acoustic signals. The Internet of Things (IoT) has provided new opportunities to gather acoustic data at lower cost and complexity through the use of battery powered wireless sensors [1]. In combination with machine learning techniques this enables rise to the vision of ubiquitously monitored industrial assets connected to real-time anomaly detection.…”

Section: Introductionmentioning

confidence: 99%

Poster Abstract: Adapting Pretrained Features for Efficient Unsupervised Acoustic Anomaly Detection

Liu

Michiels

Joosen

et al. 2022

2022 21st ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN)

Self Cite

View full text Add to dashboard Cite

Faults in industrial equipment lead to significant costs due to downtime and unplanned maintenance interventions. Acoustic Internet of Things (IoT) sensors combined with machine learning offers the possibility of early fault detection to mitigate these costs. However, prior approaches to Acoustic Anomaly Detection (AAD) are poorly suited to operating within the resource constraints of IoT systems as they require the acquisition of large volumes of data to train models from scratch for different machine types or operating environments. To overcome the limitations, we introduce a system utilizing pretrained low-dimensional features and Gaussian mixture models. Preliminary results on real-world datasets show that our proposed approach outperforms state-of-the-art solutions based on the area under the curve (AUC) score with an average of 4 different machine types. Furthermore, our approach requires far less training data, making it more suitable to operate within the power and network constraints of IoT devices.

show abstract

Section: Introductionmentioning

confidence: 99%