An Olfactory-Taste Synesthesia Model Combined With Electronic Nose and Electronic Tongue to Identify Flavor Substances

The main task of the electronic nose (e-nose) and electronic tongue (e-tongue) is to identify flavour substances. The human olfactory-taste nerve mechanisms affect the identification of flavour substances by the e-nose and e-tongue systems. Therefore, this study proposed an olfactory-taste synesthesia model built on the human olfactory-taste nerve pathways to process the e-nose and e-tongue data for flavour substance identification. Next, this study simulated the power spectral density and synchronization for the synesthesia model nodes and validated the bionic performance of the synesthesia model. The results suggested that, first, the synesthesia model characterized the dynamic characteristics of the human olfactory-taste nerve conduction pathways. Second, during stimulation, the power spectral density for the synesthesia model nodes exhibited 1/
$f$
characteristics. The synesthesia model nodes exhibited synchronization upon stimulation. The 1/
$f$
characteristics and synchronization indicated the bionic performance of the synesthesia model. Finally, this study chose beer and apple samples as flavour substances and input the olfactory and taste information for flavour substances into the synesthesia model for identification. Compared with the identification results of a single e-nose, single e-tongue, and the combined electronic nose and tongue systems under the multipattern recognition models, the synesthesia model acquired best accuracy, Kappa coefficients, and F
₁
-scores in the identification for the beer and apple samples. In conclusion, the effective identification results for flavour substances can be achieved by the synesthesia model.