A Temperature-Compensated High-Resolution Microwave Sensor Using Artificial Neural Network

In this study, a loss-compensated microwave (MW) planar sensor is used to characterize fluids at ~1 GHz. The environmental temperature is shown to adversely impact the recorded resonance frequency of the MW sensor, leading to data mixing. This issue is resolved using a feedforward artificial neural network with two hidden layers. Various concentrations of methanol in water (0%-100% with 10% increments) are measured at temperatures ranging between 22 °C and 60 °C. This smart sensor system exhibits a strong ability to discriminate the correct data regardless of erroneous interfering factors up to 92%.