Gas identification with tin oxide sensor array and self organizing maps: adaptive correction of sensor drifts

Low cost tin oxide gas sensors are inherently nonspecific. In addition they feature several non-desirable behaviors such as slow time response, nonlinearities and long term drifts. This paper shows that the combination of a gas sensor array together with self organizing maps can solve the gas classification problems. That is, the system is able to determine the gas present in the test chamber with error rates lower than 3%. The correction of the sensor drift with an adaptative SOM has also been investigated.