## Statistical Analysis of Linear Analog Circuits Using Gaussian Message Passing in Factor Graphs (2009)

### BibTeX

@MISC{Phadnis09statisticalanalysis,

author = {Miti Phadnis and Dr. Brandon and K. Eames},

title = {Statistical Analysis of Linear Analog Circuits Using Gaussian Message Passing in Factor Graphs},

year = {2009}

}

### OpenURL

### Abstract

This thesis introduces a novel application of factor graphs to the domain of analog circuits. It proposes a technique of leveraging factor graphs for performing statistical yield analysis of analog circuits that is much faster than the standard Monte Carlo/Simulation Program With Integrated Circuit Emphasis (SPICE) simulation techniques. We have designed a tool chain to model an analog circuit and its corresponding factor graph and then use a Gaussian message passing approach along the edges of the graph for yield calculation. The tool is also capable of estimating unknown parameters of the circuit given known output statistics through backward message propagation in the factor graph. The tool builds upon the concept of domain-specific modeling leveraged for modeling and interpreting different kinds of analog circuits. Generic Modeling Environment (GME) is used to design modeling environment for analog circuits. It is a configurable tool set that supports creation of domain-specific design environments for different applications. This research has developed a generalized methodology that could be applied towards design automation of different kinds of analog circuits, both linear and nonlinear. The tool has been successfully used to model linear amplifier circuits and a nonlinear Metal Oxide Semiconductor Field Effect Transistor (MOSFET) circuit. The results obtained by Monte Carlo simulationsiv performed on these circuits are used as a reference in the project to compare against the tool’s results. The tool is tested for its efficiency in terms of time and accuracy against the standard results. (104 pages) To my loving family and friends.... v vi

### Citations

1289 | Factor Graphs and the Sum-Product Algorithm - Kschischang, Frey, et al. - 2001 |

90 |
Microelectronic Circuits
- Sedra, Smith
- 1998
(Show Context)
Citation Context ...T. All of these circuits were modeled with the available blocks in the design environment of analog circuits. These cir-58 cuits follow from the definitions given in the textbook of microelectronics =-=[25]-=-. This section thoroughly discusses the plots obtained for all of the circuits mentioned above by applying the tool and Monte Carlo simulation. 4.1.1 Inverting and Non-Inverting Configurations of an A... |

40 |
Symbolic analysis methods and applications for analog circuits: a tutorial overview
- Gielen, Wambacq, et al.
- 1990
(Show Context)
Citation Context ...y. Symbolic analysis is another formal approach developed to evaluate the behavior or a specific characteristic of linear circuits in which a few or all circuit elements are represented11 as symbols =-=[7]-=-. Software is used to translate a circuit description into a symbolic expression modeling the circuit characteristic of interest. This analysis is primarily restricted to linear circuits. Algebraic (m... |

38 | The factor graph approach to model-based signal processing
- Loeliger, Dauwels, et al.
- 2007
(Show Context)
Citation Context ...a unique node for every factor and a unique edge for every variable in the graph; a factor node is connected with an edge representing some variable only if that factor is a function of that variable =-=[10]-=-. Edges connecting nodes in the graph model the factorization dependencies of the global function. Factor graphs are associated with a summary propagation algorithm that operates by passing messages a... |

30 | Factor graphs and algorithms
- Frey, Kschischang, et al.
- 1998
(Show Context)
Citation Context ...of control and signal processing systems. A factor graph is defined as a bipartite graph capable of realizing a global mathematical function as a composition of the factors of several local functions =-=[9]-=- and provides a graphical representation of such factorization. There exists a unique node for every factor and a unique edge for every variable in the graph; a factor node is connected with an edge r... |

24 |
BLADES: an artificial intelligence approach to analog circuit design
- El-Turky, Perry
- 1989
(Show Context)
Citation Context ...-sensitive state variable filter circuit is used to show the results. Some work has also been done in the development and use of expert systems for automated analog circuit design. El-Turky and Perry =-=[3]-=- discuss an automated design methodology for analog circuits that integrates formal and intuitive knowledge into one program for the design process. BLADES is a prototype design environment based on t... |

19 | SPICE: A Guide to Circuit Simulation and Analysis using PSpice - Tuinenga |

18 | On the application of factor graphs and the sum-product algorithm to ISI channels
- Colavolpe, Germi
- 2005
(Show Context)
Citation Context ...ols over an interfering channel with white Gaussian noise. The combination of factor graphs and sum product algorithm has also been applied to the development of detection algorithms for ISI channels =-=[16]-=-. The technique aims at determining aposteriori probability distributions of the transmitted symbols in the Intersymbol Interference (ISI) channels. It represents an ISI channel with the help of facto... |

12 |
Worst Case Tolerance Analysis of Linear Analog Circuits Using Sensitivity Bands
- Tian, Shi
- 2000
(Show Context)
Citation Context ...With the semiconductor industry advancing towards submicron era, it is becoming increasingly important to evaluate the tolerance of circuit performance on the parameter fluctuations2 for reliability =-=[2]-=-. Significant research has been done in this area of statistical yield analysis for analog circuits [2–7]. Most of the techniques have analyzed the impact of parameter variation on yield producing res... |

7 |
The factor graph EM algorithm: applications for LDPC codes
- Eckford
(Show Context)
Citation Context ...ed back and forth iteratively along the edges of the graph to arrive at an AR parameter estimate. The expectation Maximization algorithm has been shown as an instance of the message passing algorithm =-=[20]-=-. The sum-product algorithm is used to marginalize over the hidden random variables in the model. EM algorithm estimates unknown parameters in a system18 by calculating maximum likelihood (ML) parame... |

6 |
Statistical analysis of analog structures through variance calculation. Circuits and Systems I: Regular Papers
- Graupner, Schwarz, et al.
(Show Context)
Citation Context ...arable with standard Monte-Carlo/Simulation Program With Integrated Circuit Emphasis (SPICE)-based simulation methods in terms of accuracy while utilizing much less Central Processing Unit (CPU) time =-=[2,4,5]-=-. Monte Carlo and SPICE have been adopted as standard simulation techniques in the analog circuit community to perform statistical yield analysis of analog circuits. These techniques perform thousands... |

6 | AR model parameter estimation: from factor graphs to algorithms
- Korl, Loeliger, et al.
(Show Context)
Citation Context ...allel structure of sum-product algorithm unlike the BCJR algorithm. A technique of parameter estimation in a Gaussian auto-regressive model by message propagation in a factor graph has been discussed =-=[18]-=-. An auto-regressive (AR) model is a type of random process that is used to model different types of natural phenomena [19]. It focuses on the joint estimation of AR parameters, noise variance, and in... |

3 | PSPICE and Circuit Analysis - Keown - 1991 |

2 |
An introduction to factor graphs,” Signal Processing Magazine
- Loeliger
- 2004
(Show Context)
Citation Context ...nificance in our approach. 1.2 Factor Graphs Graphical models, like signal flow graphs, trellis diagrams, and a variety of block diagrams, have often been utilized by engineers to model their systems =-=[8]-=-. Factor graph analysis is another emerging technique in modeling theory and has been used in a variety of domains. It has the ability to model various types of control and signal processing systems. ... |

1 |
Mismatch And Noise
- Marshall, Thornton
- 2009
(Show Context)
Citation Context ...devices has been found to have an adverse impact on the digital logic schemes in several memory systems where uncertainties get introduced in the delay times, thereby causing race conditions to occur =-=[1]-=-. As the size of circuit components decrease, these effects become even more prominent. With the semiconductor industry advancing towards submicron era, it is becoming increasingly important to evalua... |

1 |
A technique for fast ac statistical analysis of analog circuits
- Rodriguez-Macias, Rodriguez-Vazquez
- 1998
(Show Context)
Citation Context ...arable with standard Monte-Carlo/Simulation Program With Integrated Circuit Emphasis (SPICE)-based simulation methods in terms of accuracy while utilizing much less Central Processing Unit (CPU) time =-=[2,4,5]-=-. Monte Carlo and SPICE have been adopted as standard simulation techniques in the analog circuit community to perform statistical yield analysis of analog circuits. These techniques perform thousands... |

1 |
Yield predictive model characterization in analog circuit design
- Ali, Wilson, et al.
- 2007
(Show Context)
Citation Context ...etability of the generated expressions. There have been efforts to design a simulation-based technique that includes yield as one of the performance parameters in the optimization process. Ali et al. =-=[6]-=- have proposed a predictive model to include yield as the cost function in the initial stages of the design and then use Monte Carlo simulation for its estimation. The two main components of the synth... |

1 |
Calibration of analog-to-digital converters with low-precision components,” in The 2007 Analog Decoding Workshop
- Loeliger
- 2007
(Show Context)
Citation Context ...tem represented by the graph through evaluation of the output for a given set of inputs by message propagation. Factor graph theory has been applied to the analysis of mixed signal circuits. Loeliger =-=[12]-=- has explored the potential of the message passing algorithm in a factor graph for the calibration of analog to digital converters made of low precision components. However, the work performed to date... |

1 |
Gme:generic modeling environment,” [http://www.isis.vanderbilt.edu/Projects/gme], 2008
- Loeliger, Hu, et al.
- 2006
(Show Context)
Citation Context ... toolset that facilitates easy creation of domain-specific modeling environments [13]. The configuration is accomplished through metamodels specifying the modeling paradigm for the application domain =-=[14]-=-. A paradigm refers to the domain language that embeds all the constructs for modeling an application within it. It contains all the information concerning syntax and semantics of the resultant modeli... |