Name
Abstract | ||
|---|---|---|
Internet of things (IoT) systems are collecting an enormous amount of sensor data which in turn play a critical role in their decision-making process. However, the more sensor data are collected, the higher probability the faulty data occurs. The existence of faulty data may drive any system to make incorrect decisions and actions. It is clear that the faulty data should be detected and corrected as earlier as possible. In other words, the sensor data should be verified in the collecting steps. To best of our knowledge, the sensor data verification in the collecting process has not been well investigated. This research proposes a framework to address this issue right before the sensor data are stored. The verification makes use of a data forecasting technique to estimate the correct range of values for each received point of sensor data. The proposed system was implemented on a real test-bed on a Raspberry Pi 3. The experimental results show the proposed system can detect and correct faulty sensor data in less than 50ms. The accuracy of the proposed method is eighty five percent on average. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1109/APCC47188.2019.9026398 | 2019 25th Asia-Pacific Conference on Communications (APCC) |
Keywords | DocType | ISSN |
Internet of Things,Data Forecasting | Conference | 2163-0771 |
ISBN | Citations | PageRank |
978-1-7281-3680-6 | 0 | 0.34 |
References | Authors | |
3 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Anh Lan Nguyen | 1 | 0 | 0.34 |
| Eiji Kamioka | 2 | 96 | 21.65 |
| Toan Nguyen-Duc | 3 | 0 | 0.34 |