Modeling of VLSI MOSFET Characteristics Using Neural Networks
Neural modeling of transistor current-voltage characteristics is explored as a possible solution to the complexity and accuracy problems currently encountered with analytical representations of VLSI devices. The neural modeling methodology is discussed along with first results obtained for a 0.8 µm CMOS process. The drain and substrate current-voltage characteristics of an n-channel MOSFET device are modeled over a drain current range of 10 orders of magnitude, from deep subthreshold to high-current operation.
KeywordsGate Voltage Neural Modeling CMOS Process Output Quantity Drain Voltage
Unable to display preview. Download preview PDF.
- Jeannette Lawrence, Introduction to Neural Networks and Expert Systems, California Scientific Software, Nevada City CA, 1992.Google Scholar
- James A. Anderson and Edward Rosenfeld, Neurocomputing: Foundations of Research, MIT Press, Cambridge MA, 1990.Google Scholar
- NeuralWare Staff, Neural Computing - NeuralWorks Professional II /PLUS and NeuralWorks Explorer, Software User’s Manual, NeuralWare Inc., Pittsburgh PA, 1991.Google Scholar
- S. Halama et al., Consistent User Interface and Task Level Architecture of a TCAD System, Proceedings NUPAD IV, Seattle WA, 1992.Google Scholar