Model-Based Parameter Estimation in Computational Electromagnetics: (Electromagnetic Waves)

Computational electromagnetics (CEM) involves modeling the interaction of electromagnetic fields with physical objects and their environment, such as the radiation emitted by antennas and the fields scattered from radar targets. First-principles or generating models (GMs) based on Maxwell's equations, provide a microscopic, spatial description of the charge and current distributions that normally require several samples per wavelength. Model-based parameter estimation (MBPE) uses a macroscopic, reduced-order, physically based fitting model (FM) to adaptively sample GM results while minimizing the number needed to quantify various EM observables such as frequency responses, far-field radiation patterns, interaction effects, etc. The FMs can reduce the needed GM sampling cost by a factor of 10 or more while yielding a continuous result of needed observables to avoid missing important details. The FMs can also indicate the numerical uncertainty of such quantities from measured as well as computed data. After an introduction to the subject and its mathematical background, subsequent chapters cover system identification, MBPE techniques and the various roles of Prony's methods as FMs in CEM. Other related topics that are covered include derivative sampling, radiation pattern synthesis and estimation, and assorted other applications. The book is aimed at the computational electromagnetics community and those working in applied sciences with complex models such as acoustics, mechanical structures, geo-physics and physics.

Table of Contents:
Chapter 1: System Identification and Model-Based Parameter Estimation Chapter 2: A Brief Sampling of System Identification and Model-Based Parameter Estimation Applications in Various Disciplines Chapter 3: Mathematical Background of MBPE Chapter 4: Sampling Strategies for Effective Implementation of Prony's Method Chapter 5: Conserving Waveform Information Content in the Spectral Domain using Prony's Method Chapter 6: Minimizing the Number of Frequency Samples Needed to Represent a Transfer Function Using Adaptive Sampling Chapter 7: Using Prony's Method to Design Arrays that Produce the Patterns of Continuous Source Distributions and Prescribed Radiation Patterns Chapter 8: Designing Discrete Arrays Using Prony's Method to Model Exponentiated Radiation Patterns Chapter 9: Using Adaptive Estimation to Minimize the Number of Samples Needed to Develop a Radiation or Scattering Pattern to a Specified Uncertainty Chapter 10: Modeling Dipole Arrays that Produce Synthesized Patterns Using NEC Chapter 11: Using Model-Based Parameter Estimation to Assess the Accuracy of Numerical Models Chapter 12: Using Prony's Method to Develop Pole-Based Models of Linear Sources Chapter 13: Inversion of One-Dimensional Scattering Data Using Prony's Method Chapter 14: Derivative Sampling of Computational Data Appendix A: MBPE estimation in computational electromagnetics Appendix B: Symbols and Notation Appendix C: MBPE References

About the Author :
Edmund K. Miller earned a PhD in Electrical Engineering from the University of Michigan in 1965 with an emphasis on computational electromagnetics. His working career has been varied including employment at four universities (Michigan Technological University, University of Michigan, Kansas University and Ohio University), three companies (MB Associates, Rockwell International Science Center, and General Research Corporation, all in California) and two national laboratories (Lawrence Livermore and Los Alamos). He was the first president of ACES.