A gas sensor system coupled to an artificial neural network capable of self-calibration against ambient humidity and temperature fluctuations

摘要

The responses of chemoresistive gas sensors suffer from the fluctuations in the background atmospheric conditions. An appropriate countermeasure is required to identify and compensate these drift-like terms introduced in the responses. Here, a non-linear model is presented for a specific Chemoresistive gas sensor, which facilitates these identification and compensation processes. Resistive gas sensor is treated as a nonlinear system that is affectedsimultaneously by three inputs: the partial pressure of the target gas, the relative humidity level and the temperature of the surrounding atmosphere, while the sensor resistance is the output. A MISO model is considered to simulate the behavior of gas sensor. Resistance of the sensor along with the relative humidity and temperature are the inputs of the model. Target gas concentration is the single output of the model. A large database was created out of the experimental results, i.e. the inputs and outputs of the system in different conditions. The model was simulated by the utilization of an appropriate artificial neural network. This was connected to the sensor and could deliver the correct contamination level upon receiving the measured gas sensor response, ambient humidity and temperature.