THE EFFECTS OF SIGNAL CONDITIONING AND QUANTIZATION UPON GAS AND ODORSENSING SYSTEM PERFORMANCE

The effects of signal conditioning and quantization during data acquis ition in multisensor arrays applied to gas and odour sensing is often given little attention. Aspects such as sensor technology and pattern recognition are often considered more important. However, if multisens or system performance is to be optimized, the signal conditioning and data acquisition must complement the sensor technology and pattern rec ognition techniques. It will be shown that system sensitivity and the sensor information content can be severely restricted by inappropriate signal conditioning. Also, the analogue to digital converter (ADC) is seen to be very important when considering the limitations of a syste m. The resolution will determine the upper limit of patterns and hence odour fingerprints that can be identified. In considering these limit ations, allowances are also be made for the concentration dependence o f the sensors upon the measurands and the pattern recognition pre-proc essing procedures. In a practical situation, the sensor reproducibilit y is usually the limiting factor when determining the level of overall system performance. This reproducibility is treated as a fractional v ariation of the ADC dynamic range, effectively reducing the quantizati on resolution. Estimates are made of the size of the multisensor array required to identify any given number of measurands.