Robust gas recognition using adaptive thresholding

We propose a simple yet robust gas recognition technique that uses adaptive thresholding to convert the outputs of a sensor array into a unique digital code. Gas recognition is achieved by simply looking for a match within a library of digital codes. The proposed technique was successfully validated using in-house tin-oxide gas sensor arrays. Experimental results show that the proposed approach mitigates the effects of sensor drift, with correct detection rates of 87.86%, 90.72%, and 95% for carbon monoxide, propane, and ethanol, respectively.