运用人工神经网络进行混合气体定量分析的研究

优选了分析CO、H2和CH4混保气体的传感器阵列，构造了传感器信号预处理和神经网络训练算法，从而建立了用于混合气体定量分析的人工嗅觉系统。实验结果证明，系统能够以较高的精度分辨出3种气体的浓度。

Quantitative Analysis of Gas Mixture Using an Artificial Neural Network

Overlapping sensitivity in gas sensors has led to great interest in developing artificial olfactory systems in recent years. Selectivity can, in principle, be improved by the use of an array of gas sensors with appropriate pattern recognition techniques. In our study, an artificial olfactory system, which is composed of a metal oxide semiconductor sensor array with a feed forward neural network, is designed to determine the individual gas concentrations of a mixture of CO, H 2 and CH 4. An improved BP algorithm is introduced in eqs. (4) (5) and (6) for training the neural network. It can not only enhance a lot the convergence rate of learning, but also lessen in some cases the difficulty of being easily trapped in local minimum. Experiments have been done by the use of a gas test station. Figs.2, 3 and 4 and Table 1, show clearly that the designed system can determine very accurately the concentrations of CO, H 2 and CH 4.