An ANN Model of Miniaturized Multi-Band-Pass Filter
Analytical Methods are normally used to solve a variety of band pass filters (BPF) antenna problems effectively and accurately, on the cost of very much time consumption. However soft computing techniques like artificial neural network (ANN) do not involve complex mathematical procedures and are also more fast and straightforward. In this article, an ANN model is proposed for analysis of miniaturized band-pass filter (BPF) with defected ground structure (DGS). The ANN model is designed and developed for prediction of length, width, location and orientation of 3 slots on ground plane with improved performance of proposed BPF for wireless application. The HFSS 16.0 (64- bit) simulation software is used to collect data samples for training and testing of ANN model. The ANN toolbox in MatLab R2013a software is used for training and testing of model. The proposed multi BPF resonates at 1.04 GHz, 1.7 GHz, 1.91 GHz, and 2.83 GHz for different wireless applications like GSM, UMTS, ISM, WiMAX, LTE Band3 etc. A prototype of proposed multi BPF is also fabricated using FR4 substrate, and its performance parameters are measured. Testing results of ANN model are observed in a very good agreement with Simulated and measured results. CAD and antenna designers can use ANN Model of proposed multi BPF for quick observation without any complexity.
Keywords:Soft Computing Technique, Artificial Neural Network, BPF, Wireless Communication, Antenna