The data set contains the recordings of 16 chemical sensors exposed to two dynamic gas mixtures at varying concentrations, acquired continuously during 12 hours.
| Associated Tasks | Classification, Regression |
| Data Source | Real |
| Dataset Characteristics | Multivariate, Time-Series |
| Date Donated | 2015-03-19 |
| Feature Type | Real |
| Labeled | Yes |
| Missing Values | No |
| Name | Gas sensor array under dynamic gas mixtures |
| Number of Features | 19 |
| Number of Instances | 4178504 |
| Source | UCI Machine Learning Repository |
| Time Series | Yes |
This data set contains the acquired time series from 16 chemical sensors exposed to gas mixtures at varying concentration levels. Two gas mixtures were generated: Ethylene and Methane in air, and Ethylene and CO in air. Each measurement was constructed by the continuous acquisition of the 16-sensor array signals for about 12 hours without interruption.
The data set was collected at the ChemoSignals Laboratory in the BioCircuits Institute, University of California San Diego. The sensor array consisted of 16 chemical sensors of 4 different types (TGS-2600, TGS-2602, TGS-2610, TGS-2620), with sensors integrated with customized signal conditioning and control electronics. Sensor conductivities were acquired continuously at 100 Hz sampling frequency within a 60 ml measurement chamber with a constant gas flow.
Concentration transitions were set randomly to simulate varying gas levels with all possible transitions included. Concentration ranges were chosen to induce similar sensor response magnitudes for Ethylene (0-20 ppm), CO (0-600 ppm), and Methane (0-300 ppm). The dataset aims to aid development of methods for continuous monitoring, sensor variability studies, sensor failure analysis, and calibration transfer research.