Premium
Non‐linear modeling of 0.18‐μM CMOS using neural network
Author(s) -
Alam M. S.,
Armstrong G. A.,
Toner B.,
Fusco V. F.
Publication year - 2003
Publication title -
microwave and optical technology letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.304
H-Index - 76
eISSN - 1098-2760
pISSN - 0895-2477
DOI - 10.1002/mop.10822
Subject(s) - transconductance , cmos , artificial neural network , extrapolation , electronic engineering , capacitance , microwave , voltage , electrical engineering , engineering , electrode , transistor , computer science , physics , telecommunications , mathematics , artificial intelligence , mathematical analysis , quantum mechanics
This paper describes an artificial‐neural‐network (ANN) based non‐linear modeling of deep‐submicron CMOS for RF applications. The neural network model is concise when compared to the conventional modeling approach based on empirical equations and can demonstrate comparable accuracy. The non‐linear voltage dependence of drain current, transconductance, and inter electrode capacitance can be characterized by a 3‐layered neural network, whose inputs are the gate‐to‐source bias voltage V gs and drain‐to‐source bias V ds . The corresponding “well‐trained” neutral network has been found to be in excellent agreement with all intrinsic parameters and the drain current with measured data exhibits good extrapolation characteristics. This model has been implemented in an Agilent ADS simulation environment and tested on a 0.18‐μm CMOS technology operating at 2.4 GHz. © 2003 Wiley Periodicals, Inc. Microwave Opt Technol Lett 37: 53–56, 2003; Published online in Wiley InterScience ( www.interscience.wiley.com ). DOI 10.1002/mop.10822