Development of C-band monopole antenna with Artificial Neural Networks Cooperation for Latex Contamination Detecting Application

Abstract

Development of a monopole antenna operating at the frequency of C band for applying in detecting contaminated latex by using the reflection coefficient (S11) is presented. The antenna is implemented on the FR4 printed circuit board which the dielectric constant is 4.1 and the thickness of the substrate is 1.414 mm. The dimension of the antenna was initially calculated with the basic theory; then was optimized in the electromagnetic simulator program. The simulated S11<-10 dB was in the range of 3.8-6.0 GHz. To simulate the contaminated latex detection, the latex was modeled by varying the dielectric constant each step by 1 from 40-48 so that there are all 9 samples. The simulation results showed that S11 increased at the frequency of 5 GHz which were -17.29 to -15.6 dB. Then, they were used to train the artificial neural networks (ANN) to find the weights value. The suitable structure of ANN consists of 10 input nodes, 20 hidden nodes, 0.001 learning rate, and the target is dielectric constant of contaminated latex. To validate the performance of ANN, it was tested and the results showed that the error rate and the accuracy were 0.15 and 97.8%, respectively. According to the decision result of ANN, it is possible to detect the contaminated latex. Therefore, the antenna was fabricated and tested. The S11<-10 dB is in the range of 3.7-6.1 GHz. The advantages of this antenna are compact size, simple construction, and easy to fabricate which are suitable for applying in detecting contaminated latex.