ABSTRACT
The rapid advancement in sensing technologies has opened great opportunities for sensor adoption in a wide range of applications. New e-nose sensors are becoming widely available and accessible. It has been applied in a wide range of applications including healthcare, air pollution control, security, and food quality control. Unfortunately, these sensors suffer from a number of issues including noise, drift, and uncertainty issues. Addressing these issues and improving the effectiveness of this technology is a critical component to their future success in real-world applications. Current e-nose systems cannot be improved without machine learning methods that can be used to draw inferences from observed patterns extracted from e-nose datasets. The overall performance of the e-nose system can be enhanced by appropriate signal processing techniques. Moreover, the recognition performance of e-nose systems needs to be enhanced with new feature extraction, pattern recognition, and machine learning methods.
