Response time based output of metal oxide gas sensors applied to evaluation of meat freshness with neural signal analysis

The resistance response to a steep change of the composition in air is inve stigated for gas sensors based on metal oxide thin films after the target o dour was injected from the chicken meat into the atmosphere of clean air. T he change of the response is described using an exponential decrease with f ew time constants for the slope. The time constants are considered as an ou tput data of the gas sensors. The output data are collected for chicken mea t from an array of the sensors. Meat freshness is evaluated basing on the s mell when the output data of the gas sensors are processed with a neural ne twork. If the time constants are used for the output, the target smell is r ecognised even with a single sensor. (C) 2000 Elsevier Science S.A. All rig hts reserved.