Bio

Research

Software

Publications

Computational Prototyping Group

MIT

Model Reduction Group

Bradley N. Bond

Keywords

modeling and simulation, nonlinear systems, model reduction, compact modeling, system stability, numerical simulation, system identification, nonlinear signal processing, interconnect modeling, analog circuit optimization

[Back to top]

Bio

I am currently a Ph.D. student in Electrical Engineering at MIT, working in the Computational Prototyping Group under Professor Luca Daniel. I am originally from State College Pennsylvania, where I received my BS in Engineering Science and Mechanics with honors in Engineering Science from Penn State University in the spring of 2004. I came to MIT in the fall of 2004, and received my SM in Electrical Engineering from MIT in the spring of 2006. In the past I have been a research intern at Cadence Research Labs and MIT Lincoln Labs.

[Back to top]

Research Interests

Broadly speaking, my reserach interests are in the modeling and simulation of dynamical systems. This includes efficient techniques for simulation of physics-based PDEs, methods for generating compact dynamical models from larger systems of equations or from input-output data, the qualitative analysis of nonlinear system properties such as stability, and efficient simulation techniques for nonlinear systems. Dynamical models describe the behavior of systems in all fields of engineering and science, and efficient modeling and simulation techniques may help facilitate, for instance, the design of analog circuits, or the analysis of biological systems.

Currently, my Ph.D. research has been towards the automatic system-level design and optimization of RF receiver chains. Specifically, I have developed several model reduction techniques for automatically generating compact models of both linear and nonlinear analog systems, such as RF inductors and power amplifiers, with an emphasis on preserving parameter dependence and system stability in the reduced models. The resulting individual compact models can then be interconnected together to allow simulation and optimization of an entire analog system, such as an RF receiver chain.

[Back to top]

Software

newton1d
A simple 1-dimensions Newton solver for Matlab. Plots each linearization as Newton progresses, a useful tool for teaching/learning about Newton's method and when it has trouble converging.

Basic Matlab circuit simulator ... COMING SOON
A Matlab circuit simulator that reads spice-like netslists and performs several types of analysis.

NLID Tool ... COMING SOON
A tool for identifying compact stable dynamical models based on input-output data. This is an implementation of the method described in my paper ??.

[Back to top]

Publications

Stable Reduced Models of Nonlinear Descriptor Systems through Piecewise Linear Approximation and Projection
IEEE Trans. on CAD of Integrated Circuits and Systems 28(10): 1467-1480 (2009)
PDF, Bibtex entry

A Log-Frequency Approach to the Identification of the Wiener-Hammerstein Model
IEEE Signal Processing Letters 16(10): 889-892 (2009)
PDF

Guaranteed Stable Model Reduction for Indefinite and Unstable Linear Systems
IEEE/ACM International Conference on Computer Aided-Design 2008: 728-735
Recipient of the IEEE/ACM William J. McCalla ICCAD Best Paper Award
PDF, Bibtex entry

Stabilizing Schemes for Piecewise-linear Reduced Order Models via Projection and Weighting Functions
IEEE/ACM International Conference on Computer-Aided Design 2007: 860-866
Nominated for the IEEE/ACM William J. McCalla ICCAD Best Paper Award
PDF, Bibtex entry

A Piecewise-Linear Moment-Matching Approach to Parameterized Model-Order Reduction for Highly Nonlinear Systems
IEEE Trans. on CAD of Integrated Circuits and Systems 26(12): 2116-2129 (2007)
PDF, Bibtex entry

Parameterized Model Order Reduction of Nonlinear Dynamical Systems
IEEE/ACM International Conference on Computer-Aided Design 2005: 487-494
PDF, Bibtex entry

[Back to top]