An active, inverse temperature modulation strategy for single sensor odorant classification

Metal oxide semiconductors are used as gas sensors and detectors. To improve their odorant discrimination capabilities their temperature can be modulated, exploring different working regimes to which gases will react differently. In this paper we present a temperature modulation algorithm that we call active and inverse. It is active because modulation depends on the online response of the sensor, and inverse because, in contrast to previous work, we first set a target value and then calculate the necessary modulation to drive the sensor to that value. Results show that this algorithm is robust and produces reliable data for odorant classification. Moreover, we successfully classify odorants in a sequence irrespective of the preceding odorant and without any clean-up procedure between them.