A Novel ANN Ensemble and Self-calibration Model in Electronic Nose for Concentration Estimation

Electronics nose (e-nose), as an artificial olfaction system, has been used in environmental monitor. This paper presents a novel concentration estimation model for improving the accuracy, robustness and stability of e-nose in long-term use. In the estimation model, two models including an ANN ensemble model and a self-calibration model are studied. The ANN ensemble model is different from single ANN that it belongs to a piecewise linearly weighted prediction model but not nonlinear prediction problem. The self-calibration model is designed for correction of the threshold network in the ensemble due to that the threshold network becomes decay which is caused by sensor drift. Experimental results demonstrate that the proposed model is very effective in real time monitoring of formaldehyde.