Name
Abstract | ||
|---|---|---|
In order to improve the operation effectiveness of an industrial manufacturing machine, it is important for the machine operator to be able to track the operation status of their machines. Unfortunately, many legacy manufacturing machines are not equipped with status monitoring capabilities. Therefore, sensor nodes such as current and vibration sensors are developed to be retrofitted on these legacy machines to enable operation status tracking. Unfortunately, there are machines; e.g., a limestone grinder, that are placed in very harsh environments where temperature and humidity prevent sensors from being directly retrofitted on them. This work aims to evaluate the viability of tracking machine operations via their sound. The use of sound recognition for operation status tracking solves the problems associated with conventional retrofitting sensors, such as limited space in a machine or placement restriction due to sensor requirements. Specifically, we evaluate the effectiveness of Mel Frequency Cepstral Coefficient (MFCC) to recognize real machine sound and infer the operation status of that machine. The experimental results are promising, showing an accuracy of 95.4 ± 0.04%. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1109/I2MTC43012.2020.9129526 | I2MTC |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Boon-Yaik Ooi | 1 | 32 | 6.14 |
| Jason Jing-Wei Lim | 2 | 0 | 0.34 |
| Wai-Kong Lee | 3 | 37 | 13.00 |
| Shervin Shirmohammadi | 4 | 1066 | 125.81 |