ISSN 1060-992X, Optical Memory and Neural Networks (Information Optics), 2011, Vol. 20, No. 4, pp. 271–278. © Allerton Press, Inc., 2011. 271 1. INTRODUCTION In recent years, high electron mobility transistors (HEMTs) have been developed with respect to low noise, low power, and MMIC applications and these are arousing increasing interest in standard processes on an industrial scale. Increasing demand for high power and low noise amplifiers in wireless communi- cation systems has motivated many new developments in solid state devices. The available devices are unable to receive signals accurately due to the vast variation in signal strength for example in satellite com- munication systems. A monolithic microwave integrated circuit (MMIC) is a microwave circuit in which the active and pas- sive components are fabricated on the same semiconductor substrate. The frequency of operation can range from 1 to over 100 GHz [1]. This wide range frequency opens the way to a wide area of applications like radar, communication satellites, wireless local area network (WLAN) and cellular telephone [2–4]. Gallium arsenide (GaAs) has been used extensively in the development of MMICs because of its suit- ability for both high frequency transistors and low loss passive components. Other substrate that can be used as substrate for MMIC are compounds from Group III and V like indium phosphate (InP), gallium phosphate (GaP) and indium arsenide (InAs) and also silicon [1]. GaAs is more dominant than others (including silicon) in MMIC because of several advantages at microwave frequency range [4]. Advances in the technology of GaAs and InP-based high electron mobility transistors (HEMTs) com- bined with cryogenic cooling have resulted in low-noise amplifiers (LNAs) with excellent performance. The performance of InP-based LNAs operated at cryogenic temperatures especially compares well with that of solid-state masers. Various applications have emerged, over the years, for systems based on InP HEMTs in fields such as radio astronomy, ground-based receivers in deep space network (DSN), automo- tive radar, atmospheric science, very wide-band communications, etc. The first two applications use cryo- genically cooled amplifiers [5]. The reliable operation of 70 nm MHEMTs on GaAs substrate and InP- based MMICs has already been demonstrated [6, 7]. ANN Modeling Approach for Designing Low Noise PHEMT Amplifier in Wireless Communication Systems P. K. Chopra a , S. Jain b , Anil Ahlawat c a HoD ECE Ajay Kumar Garg Engineering College, Ghaziabad, India e-mail: prajyot_chopra@indiatimes.com, pradeepkchopra@gmail.com b HoD-ECE, IGIT, New Delhi, India c Department of Computer Science & Engineering, Ajay Kumar Garg Engineering College, Ghaziabad, India Received June 8, 2011; in final form, September 6, 2011 Abstract—Future efficient wireless communication systems require, for high quality performance a Broadband Amplifier in the frequency range under consideration. This could be plugged into the mea- suring path so that it enables the system to perceive even the weakest signals. For this a new Scattering- parameter model for the microwave analysis of a PHEMT (Pseudomorphic High Electron Mobility Transistor) has been developed, which is valid for a wide frequency range. The developed neural net- work model is used in designing the PHEMT power amplifier. The calculated S-parameters, gain and minimum noise figure from the ANN (Artificial Neural Networks) model are the parameters which are used to design the low noise PHEMT power amplifier. The various gains so obtained from the S-parameters are plotted with the frequency and yield a closer fit to the simulated model. The neural network training has been done using Levenberg-Marqaurdt back propagation algorithm imple- mented in ANN toolbox of MATLAB software. All the results have been compared with the experi- mental data and show a close agreement and the validity of our model. Keywords: Amplifier, wireless, microwave, model. DOI: 10.3103/S1060992X11040096