The Artificial Neural Network Algorithm Based On Gas Sensor Array Design And C Language Implementation

Based on sensor arrays, the electronic nose system is of importantpractical value to gas detection and analysis. With the poor selectivity andstability of current gas sensors, it is hard to find an accurate response equationfor a gas sensor, so processing the signal of sensor arrays by artificial neuralnetwork is drawing engineers’ and technology developers’ concentrationincreasingly. This method has the following advantages: no need to establishresponse equations, greatly reducing the workload, and higher accuracy.Therefore, in order to achieve accurate detection of carbon monoxide andmethane in the mixture of gas, MQ-2, MQ-4and MQ-8semiconductor gassesnors sensitive to both carbon monoxide and methane were chosen tocomprise a gas sensor array, and their performance were tested, and the samplesfor neural network were gained. Then, a BP neural network algorithm basen onMATLAB6.5has been designed and simulated, and the performance of variousBP neural network functions has been compared. Then corresponding C++programs are designed, and an improved scheme is also put forward.