Odor assessment of automobile cabin air by machine olfaction

Odor quality in the cabin air of automobiles can be a significant factor in the decision to purchase a vehicle and the overall customer satisfaction with the vehicle over time. Current standard practice uses a human panel to rate the vehicle cabin odors on intensity, irritation, and pleasantness. However, human panels are expensive, time-consuming and complicated to administer. To address this issue, we have developed a machine olfaction approach to assess odors inside automobiles for the purpose of enhancing or replacing the human panel by evaluating the correlation between the system performance and a trained human panel. Our approach employs an ion-mobility spectrometer and a photoionization detector for measuring volatile organic compounds inside automobiles. Our olfactory system achieves good correlations (range from 0.72 to 0.84) with a trained human panel using predictive models generated by linear regression and cross-validation. Our results support the feasibility of replacing human panel assessments by a machine olfaction system.