Spectrum Enhancement: Let g denote a scalar function which is at least continuous on the surface T of a torus. Let u ≡ {u1 , u2 } denote the spatial frequency vector and G[u] the (distribution-valued) Fourier transform of g. The spectrum enhancement (SE ) algorithm, which was introduced a few years ago [e.g., [1]] consists of suitable transformations of the function H (p) [u] := |u| 2β |G[u]| 2 |a0,0 | 2 + δ[u], where δ is the Dirac measure, a0,0 is the Fourier amplitude at the origin such that F (g)[0] = a0,0 δ[u], β ∈ R + is the enhancement order and β = 2p. It has been recently shown that SE is related to Fourier transforms of derivatives of g[.], which are of integer order whenever β ∈ N [2] and of fractional order otherwise (β / ∈ N) [3]. Herewith, the role played by the enhancement order β in the fractal analysis of g[.] will be described. Fractal Radiation Patterns: Array antennas exhibiting fractal radiation patterns are synthe- sized by means of generalized Weierstraß functions [4]. For example, in one spatial dimension (x1 ) one has f [x1 ] := ∞ m=1 η(D−2)mh[ηmx1 ] where 1 < D < 2, η > 1 and h[.] is a suitably chosen bounded, periodic function of one variable. It can be shown that the box-counting (B) fractal dimension (dimB ) of the graph of f [.] (graph[f ]) satisfies dimB [graph[f ]] = D. Two-dimensional fractal radiation patterns can be synthesized in a similar way e.g., by separation of variables. Recognition of Fractal Radiation Patterns: Let simulated or measured radiation patterns be available on the whole surface of the unit sphere or part thereof. Let the pattern dimB be unknown or assume the pattern has to be classified by fractal dimension. The SE algorithm can be used to extract features from said patterns and estimate their dimB . Moreover, patterns can be classified by a procedure which includes supervised training. Some preliminary classification and recognition results will be provided, based on numerically synthesized fractal radiation patterns.

Crosta, G. (2009). Classification of Fractal Antenna Radiation Patterns by the Spectrum Enhancement Algorithm. In L. Tsang, A.B. Samokhin (a cura di), PIERS 2009 Moscow Progress In Electromagnetics Research Symposium Draft Abstracts (pp. 413-413). Cambridge, MA : The Electromagnetics Academy.

Classification of Fractal Antenna Radiation Patterns by the Spectrum Enhancement Algorithm

CROSTA, GIOVANNI FRANCO FILIPPO
2009

Abstract

Spectrum Enhancement: Let g denote a scalar function which is at least continuous on the surface T of a torus. Let u ≡ {u1 , u2 } denote the spatial frequency vector and G[u] the (distribution-valued) Fourier transform of g. The spectrum enhancement (SE ) algorithm, which was introduced a few years ago [e.g., [1]] consists of suitable transformations of the function H (p) [u] := |u| 2β |G[u]| 2 |a0,0 | 2 + δ[u], where δ is the Dirac measure, a0,0 is the Fourier amplitude at the origin such that F (g)[0] = a0,0 δ[u], β ∈ R + is the enhancement order and β = 2p. It has been recently shown that SE is related to Fourier transforms of derivatives of g[.], which are of integer order whenever β ∈ N [2] and of fractional order otherwise (β / ∈ N) [3]. Herewith, the role played by the enhancement order β in the fractal analysis of g[.] will be described. Fractal Radiation Patterns: Array antennas exhibiting fractal radiation patterns are synthe- sized by means of generalized Weierstraß functions [4]. For example, in one spatial dimension (x1 ) one has f [x1 ] := ∞ m=1 η(D−2)mh[ηmx1 ] where 1 < D < 2, η > 1 and h[.] is a suitably chosen bounded, periodic function of one variable. It can be shown that the box-counting (B) fractal dimension (dimB ) of the graph of f [.] (graph[f ]) satisfies dimB [graph[f ]] = D. Two-dimensional fractal radiation patterns can be synthesized in a similar way e.g., by separation of variables. Recognition of Fractal Radiation Patterns: Let simulated or measured radiation patterns be available on the whole surface of the unit sphere or part thereof. Let the pattern dimB be unknown or assume the pattern has to be classified by fractal dimension. The SE algorithm can be used to extract features from said patterns and estimate their dimB . Moreover, patterns can be classified by a procedure which includes supervised training. Some preliminary classification and recognition results will be provided, based on numerically synthesized fractal radiation patterns.
Breve introduzione
fractional differentiation; Weierstrass functions
English
PIERS 2009 Moscow Progress In Electromagnetics Research Symposium Draft Abstracts
Tsang, L; Samokhin, AB
18-ago-2009
The Electromagnetics Academy
413
413
Crosta, G. (2009). Classification of Fractal Antenna Radiation Patterns by the Spectrum Enhancement Algorithm. In L. Tsang, A.B. Samokhin (a cura di), PIERS 2009 Moscow Progress In Electromagnetics Research Symposium Draft Abstracts (pp. 413-413). Cambridge, MA : The Electromagnetics Academy.
open
File in questo prodotto:
File Dimensione Formato  
FractalClassifier.pdf

Accesso Aperto

Tipologia di allegato: Abstract
Dimensione 73.81 kB
Formato Adobe PDF
73.81 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/9552
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact