DocumentCode
2705644
Title
Fast modelling of microwave circuits exploiting space mapping technology
Author
Cao, Cheng ; Dai, Fei
Author_Institution
EMC Lab., Beijing Univ. of Aeronaut. & Astronaut., Beijing, China
fYear
2009
fDate
27-29 Oct. 2009
Firstpage
385
Lastpage
388
Abstract
In this paper, an efficient space mapping based surrogate modelling technique for microwave circuit is proposed. This technique ensures us to build microwave models of high accuracy and fast speed exploiting quite small amount of 3D electromagnet simulation data. This method proposed in this paper utilizes a nearest neighbour weighted methodology to build the surrogate model. The nearest neighbour technology ensures that each sample results in improved mapping between the coarse and the fine model parameters needed to execute the SM modelling. Standard space mapping modelling is enhanced by the nearest neighbour technology: (1) utilize the closest k points of the given points to gain the local information, (2)different weighted technology mimic the smooth of the parameter space. A good fitness between the coarse model response and the fine model response is demonstrated through a example of microstrip bandpass filter.
Keywords
band-pass filters; microstrip filters; microwave circuits; 3D electromagnet simulation; component space mapping; microstrip bandpass filter; microwave circuit; nearest neighbour weighted methodology; space mapping technology; surrogate modelling technique; Band pass filters; Circuit simulation; Electromagnetic modeling; Electromagnets; Microstrip filters; Microwave circuits; Microwave technology; Microwave theory and techniques; Samarium; Space technology; component; electromagnetic (EM) simulation; filter design; k nearest neighbour (K-NN); microwave circuits; parameter extraction (PE); space mapping(SM); surrogate models;
fLanguage
English
Publisher
ieee
Conference_Titel
Microwave, Antenna, Propagation and EMC Technologies for Wireless Communications, 2009 3rd IEEE International Symposium on
Conference_Location
Beijing
Print_ISBN
978-1-4244-4076-4
Type
conf
DOI
10.1109/MAPE.2009.5355732
Filename
5355732
Link To Document