Study On Optimization Of Sensor Array For E-nose System

Electronic noses based on simulation of human olfactory system have been widely used in many aspects. An electronic nose commonly has three parts, which are a sensor array, a data acquisition system and a pattern recognition method. The sensor array is the"heart"of electronic noses. And it directly affect the overall performances of electronic noses.So effectively enhancing the sensor array performance ( such as sensitivity, selectivity, stability, etc)is the key to improve the performance of electronic noses.Sensor array optimization, which uses feature selection to decide the right kinds and number of gas sensors, is one of the methods to effectively improve the performances of electronic noses. First, a volatile organic compounds (VOCs) qualitative detection experiment was carried on to test sensor array optimization method. In this experiment, the concentrations of samples are round 600ppb to 2000ppb. Through array optimization, the original sensor array which consisted of fourteen MOS gas sensors, was optimized to a nine-sensor array. The correct recognition rate under the optimized array is 100% with discriminal function analysis (DFA).The results show that array optimization could determine the optimum sensor array and the composition of the best system of sensitive materials, processing technology and working methods.On the other hand, the selectivities of metal oxide semiconductor (MOS) gas sensors are deponded on the working temperature. Through the temperature modulation, the selectivities could be improved. Most of arrays usually used the heating voltage of 5.0 V and worked at 270℃. In the VOCs qualitative detection experiment, seven different working temperatures were used. Then through array optimization, an original sensor array consisted of ten MOS gas sensors has been optimized to a four-sensor array. The DFA results of optimal features at different temperatures were compared. Results showd that it could improve selectivity (the correct recognition rate increased to 100% from 91.7%) and reduce power consumption from the traditional work temperature 270℃to the best working temperature 220℃.Another way to increase the sensor selectivities is temperature modulation. In the application of VOCs qualitative identification, sensor array worked in a set temperature modulation mode. six feature parameters: Max(ds/dt),Min(ds/dt)/minS,max(dSt/dt)/minS,max(dSt/dt)/maxS,min(dSt/dt)/maxS,min(dSt/dt), were extracted from the dynamic response curve. The original sensor array consisted of six MOS gas sensors has been optimized to a three-sensor array. And the correct recognition rate of optimized array reached 100%.The paper contains four parts: the first part details the principle and system construction of electronic nose, the significance and means of sensor array optimization, and the temperature dependence of MOS sensors. The second part presents the applition of feature selection and array optimization in VOCs qualitative detecting. The third part researches array optimization and the working temperature selection of MOS sensors.The final part describes the application of array optimization when MOS sensors worked in temperature modulation mode.